Plateau high-risk operation safety monitoring and early warning system and method based on multi-source perception

Abstract

The invention discloses a plateau high-risk operation safety monitoring and early warning system and method based on multi-source perception, and the system comprises six units: an environment parameter multi-dimensional collection and integration unit and an operation equipment state full-dimensional monitoring unit, and can collect multi-source heterogeneous data, such as plateau operation environment parameters, equipment operation states and personnel signs, in real time. After integration of the multi-source data heterogeneous fusion processing unit, the danger situation analysis unit mines potential correlation of the data, and then accurate early warning is achieved through the high-risk operation multi-source danger perception early warning unit in combination with a dynamic threshold value. The method comprises the steps of data acquisition, format conversion and fusion, deep analysis, danger judgment, early warning release and the like. According to the invention, through a multi-source data integration and optimization algorithm, the limitation of single data acquisition and insufficient analysis capability of a traditional monitoring and early warning technology is broken through, the plateau high-risk operation danger can be monitored comprehensively and dynamically, and the operation safety guarantee level is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of safety monitoring of high-risk operations in plateaus, and in particular to a safety monitoring and early warning system and method for high-risk operations in plateaus based on multi-source perception. Background Art

[0002] High-risk operations in plateau areas present significant challenges for safety monitoring and early warning due to the unique geographical environment and complex working conditions. Low air pressure, thin oxygen concentrations, and harsh climatic conditions in the plateau make operating equipment susceptible to performance degradation and frequent failures. Operators also face increased health risks due to factors such as hypoxia and hypothermia. Therefore, effective safety monitoring and early warning for high-risk operations in the plateau are crucial.

[0003] However, existing safety monitoring and early warning technologies have numerous shortcomings. For one thing, traditional monitoring systems often rely on a single data collection method, monitoring only one or a few indicators of the operating environment or equipment status, and failing to comprehensively analyze multi-source data. In high-risk plateau operations, environmental factors, equipment operating status, and personnel vital signs are interrelated and mutually influential. A single data collection method cannot fully reflect the true danger situation, leading to untimely and inaccurate warnings.

[0004] On the other hand, existing monitoring and early warning technologies have limitations in their data processing and analysis algorithms. The algorithms used by most systems cannot effectively process the large amounts of heterogeneous data generated by high-risk operations on the plateau. They struggle to identify potential correlations between data and adapt to the dynamic changes in the plateau operating environment. For example, when key parameters such as air pressure and oxygen concentration change, the early warning thresholds cannot be adjusted promptly. This causes the warning results to be out of sync with the actual danger situation and fails to provide reliable safety assurance for high-risk operations on the plateau. Summary of the Invention

[0005] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a high-risk operation safety monitoring and early warning system and method based on multi-source perception.

[0006] The technical solution adopted by the present invention is a safety monitoring and early warning system for high-risk operations in the plateau based on multi-source perception, including: a multi-dimensional acquisition integration unit for environmental parameters, which is used to collect real-time environmental parameters such as air pressure, oxygen concentration, wind speed, temperature, and humidity in high-risk operation areas on the plateau; a full-dimensional monitoring unit for the operation equipment status, which monitors the operation parameters, mechanical structure status, and electrical system parameters of the operation equipment in all directions; a multi-modal perception unit for personnel vital signs, which collects the heart rate, blood pressure, blood oxygen saturation, body temperature, and movement posture vital sign information of the operator; a multi-source data heterogeneous fusion processing unit, which integrates the collected and monitored multi-source heterogeneous data into a comprehensive system. Format conversion, feature extraction and data fusion; the dangerous situation analysis unit uses the optimized multi-head attention mechanism to conduct in-depth analysis of the fused data and explore the potential correlation between the data; the high-risk operation multi-source dangerous perception and warning unit makes a dangerous warning based on the dangerous situation analysis results and the preset thresholds. Among them, the multi-dimensional acquisition and integration unit of environmental parameters, the full-dimensional monitoring unit of the operating equipment status, and the multi-modal perception unit of personnel vital signs are respectively connected to the multi-source data heterogeneous fusion processing unit, the multi-source data heterogeneous fusion processing unit is connected to the dangerous situation analysis unit, and the dangerous situation analysis unit is connected to the multi-source dangerous perception and warning unit for high-risk operations.

[0007] Furthermore, in the dangerous situation analysis unit, the optimized multi-head attention mechanism model formula is: Among them, Q is the query matrix, which represents the feature vector of the plateau high-risk operation data that needs to be analyzed; K is the key matrix, which stores the key data features for matching; V is the value matrix, which contains the data information corresponding to the key features; d k is the dimension of the key matrix; Attention headi Represents the output of the i-th attention head; the outputs of multiple attention heads are combined to obtain the final analysis result: Where h is the number of attention heads, W O is the output weight matrix.

[0008] Furthermore, in the high-risk operation multi-source hazard perception and warning unit, the formula for constructing the high-risk operation multi-source hazard perception and warning model is: Among them, Y is the danger warning result, which takes the value of 0 or 1, 0 represents no danger, and 1 represents danger; i is the data feature of the i-th high-risk operation on the plateau after analysis based on the optimized multi-head attention mechanism; w i is the weight of the corresponding data feature; b is the bias term; σ is the activation function; combined with the special parameters of the plateau working environment, a dynamic value adjustment formula is constructed: T new =T old ×(1+α×ΔP×ΔO), where T new is the adjusted danger warning threshold, Told is the initial threshold, α is the adjustment coefficient, ΔP is the change in air pressure, and ΔO is the change in oxygen concentration.

[0009] Furthermore, the multi-dimensional acquisition integration unit of environmental parameters optimizes the environmental data fusion algorithm based on the optimized multi-head attention mechanism, and the formula is: Among them, F env is the fused environmental data feature vector, m is the number of environmental parameter types, a j is the weight of the jth environmental parameter, E j is the j-th environmental parameter data, and MHA represents the optimized multi-head attention mechanism.

[0010] Furthermore, the full-dimensional monitoring unit for the operating equipment status uses an optimized multi-head attention mechanism to build an equipment failure prediction model, the formula is: Among them, P fault is the probability of equipment failure, l is the number of equipment monitoring parameters, c k is the weight of the kth device monitoring parameter, D k is the monitoring parameter data of the kth device, and d is the bias term.

[0011] Furthermore, the personnel vital signs multimodal perception unit constructs a personnel health risk assessment model formula as follows: Among them, R health is the personnel health risk value, t is the number of personnel physical sign parameters, e s is the weight of the sth person's physical parameter, S s is the vital sign parameter data of the sth person.

[0012] Furthermore, the full-dimensional monitoring unit for the operating equipment status includes:

[0013] Equipment operating parameter monitoring subunit, which is used to monitor the speed, torque, and power operating parameters of the operating equipment. It uses sensors to collect dynamic data during the equipment operation in real time, providing basic data for equipment operation status assessment;

[0014] The equipment mechanical structure status monitoring subunit monitors the mechanical structure of the equipment, including the displacement, vibration, and wear of different components, and promptly detects abnormal changes in the mechanical structure;

[0015] The equipment electrical system parameter monitoring subunit monitors the voltage, current, and resistance parameters of the equipment electrical system to determine whether the working status of the electrical system is normal.

[0016] Furthermore, the dangerous situation analysis unit includes:

[0017] The data feature extraction subunit extracts key features from the data output by the multi-source data heterogeneous fusion processing unit, performs dimensionality reduction and feature screening on the data using a preset algorithm, and retains information valuable for risk situation analysis;

[0018] The attention weight calculation subunit calculates the attention weight of each data feature based on the extracted features and determines the importance of different data in the risk situation analysis;

[0019] The danger pattern recognition subunit identifies potential danger patterns based on the calculated attention weights and data features, providing a basis for danger warning.

[0020] Furthermore, the high-risk operation multi-source hazard perception and warning unit includes:

[0021] The warning threshold setting subunit is used to set a reasonable danger warning threshold based on the characteristics of high-risk operations in the plateau and historical data;

[0022] The early warning information generation subunit generates corresponding early warning information when the analysis result of the dangerous situation analysis unit exceeds the early warning threshold;

[0023] The early warning information release sub-unit will promptly release the generated early warning information to relevant operating personnel and management personnel so that they can take appropriate response measures.

[0024] The safety monitoring and early warning method for high-risk operations in plateaus based on multi-source perception includes the following steps:

[0025] Step S1: The environmental parameter multi-dimensional acquisition integrated unit collects the air pressure, oxygen concentration, wind speed, temperature, and humidity environmental parameters in the high-risk plateau operation area in real time. The operation equipment status full-dimensional monitoring unit monitors the operation parameters, mechanical structure status, and electrical system parameters of the operation equipment. The personnel vital sign multimodal sensing unit collects the operator's heart rate, blood pressure, blood oxygen saturation, body temperature, and movement posture vital sign information;

[0026] Step S2: The collected and monitored multi-source heterogeneous data is transmitted to a multi-source data heterogeneous fusion processing unit for format conversion, feature extraction, and data fusion processing;

[0027] Step S3: The dangerous situation analysis unit uses the optimized multi-head attention mechanism to conduct in-depth analysis on the fused data and explore potential correlations between the data;

[0028] Step S4: The high-risk operation multi-source hazard perception and warning unit makes a hazard judgment based on the hazard situation analysis results and the preset threshold value;

[0029] Step S5: If it is determined that there is a danger, the high-risk operation multi-source danger perception and warning unit generates corresponding warning information;

[0030] Step S6: Release the warning information to relevant operators and managers to complete the entire safety monitoring and warning process.

[0031] Beneficial effects: The present invention proposes a safety monitoring and early warning system and method for high-risk operations in the plateau based on multi-source perception. The system utilizes a multi-dimensional acquisition integration unit for environmental parameters, a full-dimensional monitoring unit for the operating equipment status, and a multi-modal perception unit for personnel vital signs to achieve comprehensive multi-source data collection of the high-risk operating environment, equipment, and personnel status in the plateau, changing the limitations of traditional single data collection. The multi-source data heterogeneous fusion processing unit converts the formats and extracts features of different types of data to provide a unified and effective data basis for subsequent analysis. At the data processing and analysis level, the hazard situation analysis unit deeply mines the fused data through a unique algorithm formula, which can accurately capture the potential correlations between the data. Compared with traditional algorithms, the analysis efficiency and accuracy are greatly improved when processing large amounts of heterogeneous data for high-risk operations in the plateau. The multi-source hazard perception and early warning unit for high-risk operations combines the special parameters of the plateau to construct a dynamic threshold adjustment formula, which can adjust the warning threshold in real time according to changes in air pressure, oxygen concentration, etc., to achieve accurate hazard warnings, avoiding the problem that the warning results of existing technologies are out of touch with actual hazards. This system and method form a complete link from data collection to analysis and early warning. It not only comprehensively integrates multi-source data, but also deeply processes data through optimized algorithms. It can more promptly and accurately detect potential dangers in high-risk operations on the plateau, provide reliable safety protection for operators and equipment, and significantly improve the effectiveness and reliability of safety monitoring and early warning of high-risk operations on the plateau. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a diagram of the system unit composition of the present invention;

[0033] Figure 2 The figure is a flow chart of the method steps of the present invention. DETAILED DESCRIPTION

[0034] It should be noted that, unless there is a conflict, the embodiments in this application and the features described in the embodiments can be combined with each other. The application is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] like Figure 1 As shown in the figure, the high-risk operation safety monitoring and early warning system based on multi-source sensing includes:

[0036] The integrated unit for multi-dimensional environmental parameter acquisition is used to collect real-time environmental parameters such as air pressure, oxygen concentration, wind speed, temperature, and humidity in high-risk operating areas on the plateau.

[0037] Specifically, the integrated multi-dimensional environmental parameter acquisition unit is the system's data source, responsible for real-time collection of multiple environmental parameters within high-risk plateau work areas. Equipped with a matrix of high-precision sensors, this unit simultaneously acquires key environmental indicators such as air pressure, oxygen concentration, wind speed, temperature, and humidity. The air pressure sensor uses a high-precision digital barometer with a measurement range of 300hPa to 1100hPa and an accuracy of ±0.1hPa, accurately reflecting pressure fluctuations in the plateau. The oxygen concentration sensor utilizes an electrochemical principle, with a measurement range of 0-30% Vol and an accuracy of ±0.1% Vol, enabling real-time monitoring of oxygen concentration changes and providing data support for worker health. The wind speed sensor uses a three-cup anemometer with a measurement range of 0-70m / s and an accuracy of ±0.3m / s, accurately capturing the highly variable wind conditions in the plateau. These sensors collect data at a preset frequency, which can be dynamically adjusted between 1Hz and 10Hz based on the complexity of the work environment to ensure real-time and accurate data.

[0038] The unit utilizes a distributed data collection architecture, deploying multiple sensor nodes at key locations within the operational area to form a three-dimensional monitoring network. Each sensor node transmits collected data to a data aggregation node via wired or wireless communication. The aggregation node then transmits the collected data to the multi-source heterogeneous data fusion processing unit. During communication, an encrypted transmission protocol is used to ensure the security and reliability of data transmission. The unit also features self-calibration and fault diagnosis capabilities, enabling regular sensor calibration to ensure measurement accuracy. In the event of a sensor failure, an alarm is issued, notifying maintenance personnel to replace the sensor, ensuring continuous and stable data collection.

[0039] The full-dimensional monitoring unit for the operating equipment status performs all-round monitoring of the operating parameters, mechanical structure status, and electrical system parameters of the operating equipment;

[0040] Specifically, the full-dimensional monitoring unit for operating equipment status is responsible for comprehensively monitoring the operating status of operating equipment to ensure its safe operation in the unique plateau environment. This unit monitors operating parameters, mechanical structure status, and electrical system parameters. Regarding operating parameter monitoring, sensors installed in key locations on the equipment provide real-time access to parameters such as the equipment's speed, torque, and power. The speed sensor uses a magnetoelectric speed sensor with a measurement range of 0-30,000 r / min and an accuracy of ±0.1%, enabling precise monitoring of the operating status of the equipment's rotating components. The torque sensor uses a strain gauge torque sensor with a measurement range of 0-10,000 N·m, depending on the equipment type, and an accuracy of ±0.2%, enabling real-time monitoring of the load on the equipment's transmission system. The power sensor uses an electronic power sensor, capable of accurately measuring the equipment's input and output power, with a measurement range of 0-1,000 kW and an accuracy of ±0.5%.

[0041] Regarding mechanical structure condition monitoring, displacement sensors, vibration sensors, and wear sensors are used to monitor various equipment components. The displacement sensor uses a laser displacement sensor with a measurement range of 0-50mm and an accuracy of ±0.01mm, enabling real-time monitoring of displacement changes in equipment components. The vibration sensor uses a piezoelectric vibration sensor with a measurement frequency range of 1-10,000Hz, accurately capturing equipment vibration signals and determining whether abnormal vibrations are present. The wear sensor uses an inductive wear sensor, enabling real-time monitoring of component wear, providing a basis for maintenance. Regarding electrical system parameter monitoring, voltage sensors, current sensors, and resistance sensors are used to monitor the equipment's electrical system. The voltage sensor has a measurement range of 0-1000V and an accuracy of ±0.5%; the current sensor has a measurement range of 0-1000A and an accuracy of ±0.5%; and the resistance sensor has a measurement range of 0-100kΩ and an accuracy of ±1%, enabling real-time monitoring of circuit resistance changes. Data collected by these sensors is transmitted via an industrial bus to a data processing module for analysis and processing, enabling timely identification of potential equipment failures.

[0042] The multimodal sensing unit for personnel vital signs collects the operator's heart rate, blood pressure, blood oxygen saturation, body temperature, and posture information;

[0043] Specifically, the multimodal sensing unit for personnel vital signs focuses on collecting physiological and behavioral characteristic data from workers, providing a basis for personnel safety assurance. This unit combines wearable devices with fixed monitoring equipment to achieve real-time collection of multi-dimensional vital sign information such as a worker's heart rate, blood pressure, blood oxygen saturation, body temperature, and movement posture. Wearable devices include smart bracelets and smart helmets. Smart bracelets integrate multiple sensors, including heart rate sensors and accelerometers, to monitor a worker's heart rate, number of steps, and exercise status in real time. The heart rate sensor uses a photoelectric heart rate sensor with a measurement range of 30-220 beats / minute and an accuracy of ±1 beat / minute. The accelerometer uses a three-axis accelerometer, which can monitor a worker's movement acceleration and posture changes in real time, with a measurement range of ±16g and an accuracy of ±0.001g. The smart helmet integrates a blood oxygen saturation sensor and a temperature sensor. The blood oxygen saturation sensor uses a reflective blood oxygen sensor with a measurement range of 70-100% and an accuracy of ±2%. It can monitor changes in people's blood oxygen saturation in real time; the temperature sensor uses a digital temperature sensor with a measurement range of -40℃ to +125℃ and an accuracy of ±0.1℃. It can monitor people's body surface temperature in real time.

[0044] Fixed monitoring equipment, including infrared thermal imagers and video surveillance systems, is installed at key locations within the work area, such as rest areas and work stations. The infrared thermal imagers monitor personnel's body temperature distribution in real time, with a measurement range of -20°C to +550°C and an accuracy of ±0.5°C, enabling rapid detection of abnormal body temperatures. The video surveillance system, utilizing high-definition cameras and intelligent analysis algorithms, monitors personnel's movements, posture, and behavior in real time, identifying any dangerous behavior. Data collected by each monitoring device is transmitted wirelessly to a data processing center for comprehensive analysis and processing. The system can promptly detect health anomalies based on changes in personnel's vital signs and issue early warnings to ensure the safety of workers.

[0045] The multi-source data heterogeneous fusion processing unit performs format conversion, feature extraction and data fusion on the multi-source heterogeneous data collected and monitored above;

[0046] Specifically, the multi-source heterogeneous data fusion processing unit is the system's data hub, responsible for integrating and processing heterogeneous data from different data sources. This unit utilizes a layered processing architecture, comprising a data preprocessing layer, a feature extraction layer, and a data fusion layer. In the data preprocessing layer, the collected raw data is first formatted, converting the different formats of data collected by different sensors into a standardized data format. Simultaneously, the data is denoised using algorithms such as median filtering and Kalman filtering to remove noise interference and improve data quality. During data transmission, a data compression algorithm is used to compress the data, reducing the amount of data transmitted and improving transmission efficiency.

[0047] At the feature extraction layer, various feature extraction algorithms are used to extract key features from preprocessed data. For environmental parameter data, features such as the rate of change of air pressure and the trend of oxygen concentration changes are extracted; for equipment status data, features such as speed fluctuation and vibration frequency are extracted; and for personnel vital sign data, features such as heart rate variability and blood oxygen saturation change are extracted. Feature extraction reduces data dimensionality and highlights important information within the data. At the data fusion layer, fusion algorithms based on Bayesian networks and DS evidence theory are used to fuse the extracted features. This layer assigns appropriate weights to each data point based on its importance and credibility, organically integrating multi-source data to form a unified feature vector, providing comprehensive and accurate data support for subsequent hazard situation analysis.

[0048] The dangerous situation analysis unit uses the optimized multi-head attention mechanism to conduct in-depth analysis of the fused data and explore potential correlations between the data;

[0049] Specifically, the dangerous situation analysis unit is the core analysis module of the system, responsible for conducting in-depth analysis of the fused data and exploring potential connections between the data. This unit uses an improved multi-head attention mechanism, which uses multiple attention heads to simultaneously focus on different aspects of the data and capture complex relationships in the data. During implementation, the fused data is first feature encoded and converted into a high-dimensional vector representation. Then, through operations on the query matrix, key matrix, and value matrix, the similarity between the data is calculated to determine the importance of different data in the analysis. The optimized multi-head attention mechanism introduces technologies such as adaptive learning rate and residual connection to improve the learning and generalization capabilities of the model.

[0050] This unit divides data into time series, analyzes the data within each time window, and identifies patterns and trends in the data. By comparing and analyzing historical and real-time data, it detects abnormal changes in the data and determines whether there are potential dangers. At the same time, the unit also considers the impact of high-altitude environmental factors on dangerous situations, such as the impact of factors such as changes in air pressure and reduced oxygen concentration on equipment performance and personnel health. By establishing a correlation model between environmental factors and dangerous situations, it can accurately assess and predict dangerous situations in high-risk operations on the plateau, providing a basis for subsequent danger warnings.

[0051] The multi-source hazard perception and warning unit for high-risk operations issues hazard warnings based on the results of hazard situation analysis and preset thresholds.

[0052] Specifically, the high-risk operations multi-source hazard perception and warning unit is the system's output module, responsible for issuing hazard warnings based on the results of hazard situation analysis. This unit employs a tiered warning mechanism, categorizing warning levels into level one, level two, and level three, depending on the degree of hazard. Regarding warning threshold setting, the characteristics of high-risk operations on the plateau and historical data are combined to establish corresponding thresholds for different types of hazards. Furthermore, considering the impact of plateau environmental factors, a dynamic threshold adjustment strategy is implemented, adjusting warning thresholds in real time based on changes in environmental parameters such as air pressure and oxygen concentration to improve the accuracy of warnings.

[0053] When the hazard analysis results exceed the warning threshold, the Warning Information Generation Subunit immediately generates corresponding warning information, including detailed information such as the warning level, hazard type, occurrence time, and location. The Warning Information Dissemination Subunit promptly disseminates warning information to relevant operators and managers through various means, including audio and visual alarms, SMS notifications, and app push notifications, ensuring timely access to warning information. This unit also features warning information recording and query capabilities, recording and storing historical warning information to facilitate subsequent analysis and summary, providing a basis for improving safety management measures.

[0054] The multi-dimensional collection integration unit of environmental parameters, the full-dimensional monitoring unit of operating equipment status, and the multi-modal perception unit of personnel vital signs are respectively connected to the multi-source data heterogeneous fusion processing unit, the multi-source data heterogeneous fusion processing unit is connected to the hazard situation analysis unit, and the hazard situation analysis unit is connected to the multi-source hazard perception and warning unit for high-risk operations.

[0055] Preferably, in the dangerous situation analysis unit, the optimized multi-head attention mechanism model formula is: Among them, Q is the query matrix, which represents the feature vector of the plateau high-risk operation data that needs to be analyzed; K is the key matrix, which stores the key data features for matching; V is the value matrix, which contains the data information corresponding to the key features; d k is the dimension of the key matrix; Attention headi Represents the output of the i-th attention head. Combining the outputs of multiple attention heads yields the final analysis result: Where h is the number of attention heads, W O In order to output the weight matrix, the above formula is used to conduct multi-dimensional correlation analysis on multi-source data of high-risk operations in the plateau, and to explore the hidden dangerous situation characteristics between the data.

[0056] Specifically, the hazard situation analysis unit simulates human attention allocation patterns by constructing specific computational logic to conduct in-depth analysis of the fused multi-source data. During implementation, the multi-source data is first converted into a matrix form suitable for calculation. By operating on the query matrix, key matrix, and value matrix, the similarity between the data features is calculated to determine the importance of different data in the hazard situation analysis. By analyzing data from different angles through multiple attention heads and combining the output results of each attention head, it is possible to fully explore the complex hidden connections between the data, achieve in-depth analysis of the hazard situation of high-risk operations on the plateau, and provide accurate data support for subsequent warnings.

[0057] Preferably, in the high-risk operation multi-source hazard perception and warning unit, the formula for constructing the high-risk operation multi-source hazard perception and warning model is: Among them, Y is the danger warning result, which takes the value of 0 or 1, 0 represents no danger, and 1 represents danger; i is the data feature of the i-th high-risk operation on the plateau after analysis based on the optimized multi-head attention mechanism; w i is the weight of the corresponding data feature; b is the bias term; σ is the activation function. At the same time, the dynamic value adjustment formula is constructed based on the special parameters of the plateau operating environment: T new =T old ×(1+α×ΔP×ΔO), where T new is the adjusted danger warning threshold, T old is the initial threshold, α is the adjustment coefficient, ΔP is the change in air pressure, and ΔO is the change in oxygen concentration. The above formula can be used to achieve accurate perception and dynamic warning of the dangers of high-risk operations in the plateau.

[0058] Specifically, the multi-source hazard perception and warning unit for high-risk operations quantifies hazards by establishing a mathematical model. First, based on the data features analyzed by the optimized multi-head attention mechanism, a weighted sum is performed according to the corresponding weights and bias terms. Then, through activation function processing, a hazard warning result is obtained to clarify whether the current working environment is dangerous. At the same time, considering the particularities of the plateau environment, the initial warning threshold is dynamically adjusted based on changes in key environmental parameters such as air pressure and oxygen concentration. This ensures that the warning threshold can adapt to changes in the plateau environment in real time, improves the accuracy and timeliness of hazard perception, and achieves accurate early warning of dangers in high-risk operations on the plateau.

[0059] Preferably, the environmental parameter multi-dimensional acquisition integration unit optimizes the environmental data fusion algorithm based on the optimized multi-head attention mechanism, and the formula is: Among them, F env is the fused environmental data feature vector, m is the number of environmental parameter types, a j is the weight of the jth environmental parameter, E jis the jth environmental parameter data, and MHA represents the optimized multi-head attention mechanism. For the air pressure, oxygen concentration, and wind speed parameters in the plateau environment, this formula is used to perform correlation analysis and feature fusion on each parameter, resulting in a more comprehensive and accurate representation of the environmental data features.

[0060] Specifically, the multi-dimensional environmental parameter acquisition integration unit utilizes an optimized algorithm to perform a fusion analysis of multiple environmental parameters in plateau environments. During implementation, each environmental parameter is assigned a corresponding weight, and based on an optimized multi-head attention mechanism, correlation analysis and feature fusion are performed on environmental parameter data such as air pressure, oxygen concentration, and wind speed. This approach fully considers the interrelationships between various environmental parameters, avoiding the limitations of single-parameter analysis. This results in a more comprehensive and accurate representation of environmental data features, providing a higher-quality data foundation for subsequent data-based hazard analysis.

[0061] Preferably, the full-dimensional monitoring unit for the operating equipment status uses an optimized multi-head attention mechanism to build an equipment failure prediction model, the formula is: Among them, P fault is the probability of equipment failure, l is the number of equipment monitoring parameters, c k is the weight of the kth device monitoring parameter, D k is the monitoring parameter data of the kth device, and d is the bias term. This model combines the parameter changes of equipment operation under special plateau conditions and uses an optimized multi-head attention mechanism to conduct in-depth analysis of equipment status data to predict the possibility of equipment failure.

[0062] Specifically, the full-dimensional equipment status monitoring unit leverages a constructed equipment failure prediction model and analyzes equipment operating parameters in light of the unique plateau operating conditions. During implementation, the optimized multi-head attention mechanism deeply processes equipment operating parameter data, exploring potential connections and patterns of change between parameters. A specific calculation method is then used to derive the probability of equipment failure. This model analyzes equipment operating status data in real time, identifying potential equipment failures in advance. This provides a scientific basis for equipment maintenance and fault prevention, ensuring the safe and stable operation of equipment in plateau environments.

[0063] Preferably, the personnel vital signs multimodal sensing unit constructs a personnel health risk assessment model using the formula: Among them, R health is the personnel health risk value, t is the number of personnel physical sign parameters, e s is the weight of the sth person's physical parameter, S s is the vital sign parameter data of the sth person. Considering the impact of the plateau environment on human physiological indicators, an optimized multi-head attention mechanism is used to combine and analyze multiple vital sign data of personnel to assess their health risk status.

[0064] Specifically, the personnel health risk assessment model constructed by the multimodal perception unit for personnel vital signs comprehensively considers multiple personnel vital sign parameters and the impact of the plateau environment on human physiological indicators. By assigning corresponding weights to different vital sign parameters and utilizing an optimized multi-head attention mechanism to comprehensively analyze vital sign data such as heart rate, blood pressure, and blood oxygen saturation, the impact of each vital sign parameter is quantitatively assessed to derive a personnel health risk value. This model can monitor personnel health in real time and promptly detect health anomalies, providing strong support for protecting the health of personnel working in high-risk areas on the plateau.

[0065] Preferably, the full-dimensional monitoring unit for the operating equipment status includes:

[0066] Equipment operating parameter monitoring subunit, which is used to monitor the speed, torque, and power operating parameters of the operating equipment. It uses sensors to collect dynamic data during the equipment operation in real time, providing basic data for equipment operation status assessment;

[0067] The equipment mechanical structure status monitoring subunit monitors the mechanical structure of the equipment, including the displacement, vibration, and wear of different components, and promptly detects abnormal changes in the mechanical structure;

[0068] The equipment electrical system parameter monitoring subunit monitors the voltage, current, and resistance parameters of the equipment electrical system to determine whether the working status of the electrical system is normal.

[0069] Specifically, the three subunits of the full-dimensional monitoring unit for the operating equipment status work together to achieve comprehensive monitoring of the operating equipment. The equipment operating parameter monitoring subunit collects dynamic operating parameters such as speed, torque, and power in real time through sensors installed in key parts of the equipment, providing data for the basic assessment of the equipment's operating status; the equipment mechanical structure status monitoring subunit uses displacement, vibration, wear and other sensors to monitor the physical status of different parts of the equipment and promptly detect abnormal changes in the mechanical structure; the equipment electrical system parameter monitoring subunit uses voltage, current, resistance and other sensors to monitor the operating parameters of the equipment's electrical system and determine whether the electrical system is operating normally. The data from the three subunits together constitute a complete understanding of the operating equipment status.

[0070] Preferably, the dangerous situation analysis unit includes:

[0071] The data feature extraction subunit extracts key features from the data output by the multi-source data heterogeneous fusion processing unit, performs dimensionality reduction and feature screening on the data using a preset algorithm, and retains information valuable for risk situation analysis;

[0072] The attention weight calculation subunit calculates the attention weight of each data feature based on the extracted features and determines the importance of different data in the risk situation analysis;

[0073] The danger pattern recognition subunit identifies potential danger patterns based on the calculated attention weights and data features, providing a basis for danger warning.

[0074] Specifically, the three subunits of the Hazard Situation Analysis Unit form a complete hazard situation analysis process. The Data Feature Extraction Subunit uses a specific algorithm to filter out key features valuable for hazard situation analysis from the fused multi-source data, reducing the data dimension. The Attention Weight Calculation Subunit calculates the importance of each data feature in the hazard situation analysis based on the extracted features. The Hazard Pattern Recognition Subunit identifies potential hazard patterns based on the calculated attention weights and data features, thus providing accurate hazard judgments for the High-Risk Operation Multi-Source Hazard Perception and Warning Unit, enabling effective analysis of the hazard situation of high-risk operations on the plateau.

[0075] Preferably, the high-risk operation multi-source hazard perception and warning unit includes:

[0076] The warning threshold setting subunit is used to set a reasonable danger warning threshold based on the characteristics of high-risk operations in the plateau and historical data;

[0077] The early warning information generation subunit generates corresponding early warning information when the analysis result of the dangerous situation analysis unit exceeds the early warning threshold;

[0078] The early warning information release sub-unit will promptly release the generated early warning information to relevant operating personnel and management personnel so that they can take appropriate response measures.

[0079] Specifically, the three subunits of the high-risk operations multi-source hazard perception and warning unit carry out hazard warning work in an orderly manner. The warning threshold setting subunit sets a reasonable initial hazard warning threshold based on historical data and operational characteristics of high-risk operations on the plateau. The warning information generation subunit generates warning information containing detailed information such as hazard type and occurrence time according to established rules when the hazard situation analysis results exceed the preset threshold. The warning information dissemination subunit promptly transmits the generated warning information to operators and management personnel through various communication methods, ensuring that relevant personnel can quickly obtain hazard information and take countermeasures to ensure the safety of high-risk operations on the plateau.

[0080] like Figure 2 As shown in FIG, the safety monitoring and early warning method for high-risk operations in plateaus based on multi-source perception includes the following steps:

[0081] Step S1: The environmental parameter multi-dimensional acquisition integrated unit collects the air pressure, oxygen concentration, wind speed, temperature, and humidity environmental parameters in the high-risk plateau operation area in real time. The operation equipment status full-dimensional monitoring unit monitors the operation parameters, mechanical structure status, and electrical system parameters of the operation equipment. The personnel vital sign multimodal sensing unit collects the operator's heart rate, blood pressure, blood oxygen saturation, body temperature, and movement posture vital sign information;

[0082] Step S2: The collected and monitored multi-source heterogeneous data is transmitted to a multi-source data heterogeneous fusion processing unit for format conversion, feature extraction, and data fusion processing;

[0083] Step S3: The dangerous situation analysis unit uses the optimized multi-head attention mechanism to conduct in-depth analysis on the fused data and explore potential correlations between the data;

[0084] Step S4: The high-risk operation multi-source hazard perception and warning unit makes a hazard judgment based on the hazard situation analysis results and the preset threshold value;

[0085] Step S5: If it is determined that there is a danger, the high-risk operation multi-source danger perception and warning unit generates corresponding warning information;

[0086] Step S6: Release the warning information to relevant operators and managers to complete the entire safety monitoring and warning process.

[0087] In terms of data collection, traditional technologies are often limited to a single type of data and cannot fully reflect the complex conditions of the work site. This system, with its multi-dimensional environmental parameter acquisition integration unit, full-dimensional operating equipment status monitoring unit, and multimodal personnel vital sign sensing unit, enables the simultaneous collection of multi-source data such as environmental parameters, equipment operating status, and personnel vital signs. Whether it is changes in air pressure and oxygen concentration unique to plateau regions, subtle displacements of equipment mechanical structures, or heart rate fluctuations of personnel, these can all be accurately captured, effectively addressing the flaw of one-sided data and providing a rich and comprehensive data foundation for subsequent analysis.

[0088] In the data processing and analysis phase, traditional algorithms have difficulty in exploring the potential connections between multi-source data, and are unable to adapt to the changing working environment of the plateau. This system introduces an optimized multi-head attention mechanism that can deeply analyze the fused multi-source data. This mechanism simulates the way humans allocate attention, focusing on the key and valuable parts of the data, analyzing the inherent connections between the data from different angles, and accurately identifying potential dangerous patterns. At the same time, the high-risk operation multi-source hazard perception and warning unit combines the special parameters of the plateau working environment to establish a dynamic threshold adjustment mechanism. According to changes in key environmental factors such as air pressure and oxygen concentration, the hazard warning threshold is calibrated in real time, changing the inflexible and inaccurate situation of traditional fixed threshold warnings, and significantly improving the timeliness and accuracy of warnings.

[0089] In addition, this system adopts a modular design and collaborative operation mode, with each unit having a clear division of labor and close cooperation. The multi-source data heterogeneous fusion processing unit integrates data in different formats to eliminate data barriers; the sub-units of the full-dimensional monitoring unit of the operating equipment status ensure equipment safety monitoring from multiple dimensions such as operating parameters, mechanical structure, and electrical system; the high-risk operation multi-source hazard perception and early warning unit forms a complete early warning process through threshold setting, information generation and release sub-units. This systematic architectural design enables the entire system to operate efficiently, not only overcoming the shortcomings of existing technologies, but also forming a complete closed loop from data collection, fusion, analysis to early warning, providing a reliable and comprehensive safety assurance system for high-risk operations on the plateau, and greatly improving the level of operational safety management.

[0090] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," "connected," and "fixed" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0091] While embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A safety monitoring and early warning system for high-risk operations in plateaus based on multi-source perception, characterized by: include: The integrated unit for multi-dimensional environmental parameter acquisition is used to collect real-time environmental parameters such as air pressure, oxygen concentration, wind speed, temperature, and humidity in high-risk operating areas on the plateau. The full-dimensional monitoring unit for the operating equipment status performs all-round monitoring of the operating parameters, mechanical structure status, and electrical system parameters of the operating equipment; The multimodal perception unit of personnel vital signs collects the vital sign information of the workers, such as heart rate, blood pressure, blood oxygen saturation, body temperature, and movement posture; the multi-source data heterogeneous fusion processing unit converts the format, extracts features and fuses the collected and monitored multi-source heterogeneous data; the dangerous situation analysis unit uses the optimized multi-head attention mechanism to conduct in-depth analysis of the fused data and explore the potential correlation between the data; the multi-source dangerous perception and early warning unit for high-risk operations issues dangerous warnings based on the dangerous situation analysis results and preset thresholds. Among them, the multi-dimensional acquisition and integration unit of environmental parameters, the full-dimensional monitoring unit of the operating equipment status, and the multi-modal perception unit of personnel vital signs are respectively connected to the multi-source data heterogeneous fusion processing unit, the multi-source data heterogeneous fusion processing unit is connected to the dangerous situation analysis unit, and the dangerous situation analysis unit is connected to the multi-source dangerous perception and early warning unit for high-risk operations.

2. The high-risk operation safety monitoring and early warning system based on multi-source perception according to claim 1 is characterized in that: In the dangerous situation analysis unit, the optimized multi-head attention mechanism model formula is: Among them, Q is the query matrix, which represents the feature vector of the plateau high-risk operation data that needs to be analyzed; K is the key matrix, which stores the key data features for matching; V is the value matrix, which contains the data information corresponding to the key features; d k is the dimension of the key matrix; Attention headi Represents the output of the i-th attention head; the outputs of multiple attention heads are combined to obtain the final analysis result: Where h is the number of attention heads, W O is the output weight matrix.

3. The high-risk operation safety monitoring and early warning system based on multi-source perception according to claim 1 is characterized in that: In the high-risk operation multi-source hazard perception and warning unit, the formula for constructing the high-risk operation multi-source hazard perception and warning model is: Among them, Y is the danger warning result, which takes the value of 0 or 1, 0 represents no danger, and 1 represents danger; i is the data feature of the i-th high-risk operation on the plateau after analysis based on the optimized multi-head attention mechanism; w i is the weight of the corresponding data feature; b is the bias term; σ is the activation function; combined with the special parameters of the plateau working environment, a dynamic value adjustment formula is constructed: T new =T old ×(1+α×ΔP×ΔO), where T new is the adjusted danger warning threshold, T old is the initial threshold, α is the adjustment coefficient, ΔP is the change in air pressure, and ΔO is the change in oxygen concentration.

4. The multi-source sensing-based high-risk operation safety monitoring and early warning system for plateaus according to claim 1 is characterized in that: The multi-dimensional acquisition integration unit of environmental parameters optimizes the environmental data fusion algorithm based on the multi-head attention mechanism. The formula is: Among them, F env is the fused environmental data feature vector, m is the number of environmental parameter types, a j is the weight of the jth environmental parameter, E j is the j-th environmental parameter data, and MHA represents the optimized multi-head attention mechanism.

5. The high-risk operation safety monitoring and early warning system based on multi-source perception according to claim 1 is characterized in that: The full-dimensional monitoring unit for the operating equipment status uses an optimized multi-head attention mechanism to build an equipment failure prediction model, the formula is: Among them, P fault is the probability of equipment failure, l is the number of equipment monitoring parameters, c k is the weight of the kth device monitoring parameter, D k is the monitoring parameter data of the kth device, and d is the bias term.

6. The high-risk operation safety monitoring and early warning system based on multi-source sensing according to claim 1 is characterized in that: The personnel vital signs multimodal sensing unit constructs a personnel health risk assessment model with the formula: Among them, R health is the personnel health risk value, t is the number of personnel physical sign parameters, e s is the weight of the sth person's physical parameter, S s is the vital sign parameter data of the sth person.

7. The multi-source sensing-based high-risk operation safety monitoring and early warning system for plateaus according to claim 1 is characterized in that: The full-dimensional monitoring unit for the operating equipment status includes: Equipment operating parameter monitoring subunit, which is used to monitor the speed, torque, and power operating parameters of the operating equipment. It uses sensors to collect dynamic data during the equipment operation in real time, providing basic data for equipment operation status assessment; The equipment mechanical structure status monitoring subunit monitors the mechanical structure of the equipment, including the displacement, vibration, and wear of different components, and promptly detects abnormal changes in the mechanical structure; The equipment electrical system parameter monitoring subunit monitors the voltage, current, and resistance parameters of the equipment electrical system to determine whether the working status of the electrical system is normal.

8. The high-risk operation safety monitoring and early warning system based on multi-source sensing according to claim 1 is characterized in that: The dangerous situation analysis unit includes: The data feature extraction subunit extracts key features from the data output by the multi-source data heterogeneous fusion processing unit and performs dimensionality reduction and feature screening on the data using a preset algorithm; The attention weight calculation subunit calculates the attention weight of each data feature based on the extracted features and determines the importance of different data in the risk situation analysis; The danger pattern recognition subunit identifies potential danger patterns based on the calculated attention weights and data features, providing a basis for danger warning.

9. The high-risk operation safety monitoring and early warning system based on multi-source sensing according to claim 1 is characterized in that: The high-risk operation multi-source hazard perception and warning unit includes: The warning threshold setting subunit is used to set a reasonable danger warning threshold based on the characteristics of high-risk operations in the plateau and historical data; The early warning information generation subunit generates corresponding early warning information when the analysis result of the dangerous situation analysis unit exceeds the early warning threshold; The early warning information release sub-unit will release the generated early warning information to relevant operating personnel and management personnel in a timely manner.

10. A safety monitoring and early warning method for high-risk operations in plateaus based on multi-source perception is characterized by: The following steps are involved: Step S1: The environmental parameter multi-dimensional acquisition integrated unit collects the air pressure, oxygen concentration, wind speed, temperature, and humidity environmental parameters in the high-risk plateau operation area in real time. The operation equipment status full-dimensional monitoring unit monitors the operation parameters, mechanical structure status, and electrical system parameters of the operation equipment. The personnel vital sign multimodal sensing unit collects the operator's heart rate, blood pressure, blood oxygen saturation, body temperature, and movement posture vital sign information; Step S2: The collected and monitored multi-source heterogeneous data is transmitted to a multi-source data heterogeneous fusion processing unit for format conversion, feature extraction, and data fusion processing; Step S3: The dangerous situation analysis unit uses the optimized multi-head attention mechanism to conduct in-depth analysis on the fused data and explore potential correlations between the data; Step S4: The high-risk operation multi-source hazard perception and warning unit makes a hazard judgment based on the hazard situation analysis results and the preset threshold value; Step S5: If it is determined that there is a danger, the high-risk operation multi-source danger perception and warning unit generates corresponding warning information; Step S6: Release the warning information to relevant operators and managers to complete the entire safety monitoring and warning process.

Images