Dangerous engineering safety management system

Abstract

The invention discloses a dangerous and large project safety management system, and belongs to the technical field of building construction safety monitoring. According to the system, vibration sensors are arranged on the boundary of a construction area, inclination sensors are installed on the surface of a supporting structure, static force level gauges are arranged corresponding to settlement monitoring points, and soil pressure boxes are buried in the side wall of a foundation pit in a layered mode, so that multi-dimensional data collection is achieved. A finite element analysis model is adopted in an edge calculation unit arranged in the area controller, supporting structure material parameters serve as input, and theoretical values are dynamically calculated in combination with real-time monitoring data. When the deviation between the measured data and the theoretical value exceeds a first threshold value, the historical mean value is automatically fused for iterative calculation, and if the deviation still exceeds a second threshold value after continuous three times of iteration, primary early warning is triggered. The system can evaluate the safety state of the supporting structure in real time, effectively prevents collapse accidents, and is suitable for safety management of dangerous and large projects such as building foundation pits and tunnels.

Description

Technical Field

[0001] The present invention relates to the field of construction safety management, and more particularly to a safety management system for dangerous and major projects. Background Art

[0002] In the construction industry, hazardous projects (highly dangerous sub-projects) refer to those projects that occur during the construction process and could potentially result in mass casualties among workers or cause significant adverse social impacts. According to relevant regulations of the Ministry of Housing and Urban-Rural Development, these projects include foundation pit projects, formwork projects and support systems, lifting and hoisting, and the installation and disassembly of lifting machinery. Deep foundation pit projects, as the most common type of hazardous project, are becoming increasingly widespread with the acceleration of urbanization and the continuous expansion of building scale. For example, in the construction of large-scale commercial complexes and high-rise and super-high-rise buildings in cities, deep foundation pits provide the necessary space for underground structure construction. However, the safety management of deep foundation pit projects faces many severe challenges, especially the difficulty in warning of collapse accidents caused by structural instability in deep foundation pits.

[0003] Most deep foundation pit collapse accidents occur when the deformation monitoring of the supporting structure fails. For dangerous projects such as deep foundation pits, the traditional safety management system has the following obvious defects: 1. The problem of insufficient coordination of multi-parameter monitoring. The existing safety monitoring system usually arranges vibration monitoring, tilt monitoring, settlement monitoring and soil pressure monitoring modules independently. The discrete monitoring leads to poor data synchronization, which ultimately leads to delayed warnings and causes local collapse accidents; 2. The problem of lagging real-time analysis capabilities. The traditional finite element analysis model needs to import monitoring data offline, and the completion of meshing, parameter setting, solution calculation and other processes takes a long time. In the case of a race against time in deep foundation pit construction, long analysis delays make The problem of the mechanical response of the support structure cannot be grasped in time; 3. The problem of low intelligence of the early warning mechanism. The existing safety management system mostly adopts fixed threshold alarms, and does not consider the dynamic characteristics of the structural mechanical response. Faced with complex and changeable construction environments and geological conditions, the static early warning mechanism is difficult to adapt to changes in actual conditions; 4. The problem of weak data fusion capabilities. There is a lack of effective fusion methods for different types of sensor data. For example, there is a correlation between soil pressure and vibration intensity. Because the correlation between data is ignored, the early warning signal cannot be accurately identified. A large amount of valuable monitoring data cannot be effectively integrated and analyzed, and it is difficult to play its due role, and it is impossible to provide a comprehensive and accurate decision-making basis for the safety management of dangerous projects.

[0004] How to build a safety management system for dangerous and major projects that can achieve multi-parameter coordinated monitoring, real-time analysis capabilities, intelligent early warning and efficient data fusion has become an urgent problem to be solved. Summary of the Invention

[0005] An object of the present invention is to solve at least the above problems and to provide at least the advantages which will be described hereinafter.

[0006] Another purpose of the present invention is to solve the problem that traditional dangerous engineering monitoring systems are difficult to assess the safety status of foundation pits in real time, comprehensively and accurately, and are unable to effectively integrate multi-source data for dynamic analysis and early warning, resulting in difficulty in detecting potential risks of foundation pits in advance.

[0007] Another purpose of the present invention is to solve the problem of how to reasonably fuse current monitoring data with the mean of historical data to improve the scientificity and accuracy of monitoring data processing, and at the same time set appropriate fusion parameters according to the characteristics of different sensors.

[0008] Another purpose of the present invention is to solve the problem that abnormal data interferes with the accuracy of monitoring results during the monitoring process. It is necessary to establish an effective mechanism to filter abnormal data, ensure the reliability of data input into the finite element analysis model, and improve the credibility of the analysis results.

[0009] Another purpose of the present invention is to solve the problem that the accuracy of the sensor may decrease as the monitoring progresses. A mechanism is needed to trigger the sensor calibration in time based on the data fusion calculation results to ensure the long-term stability and accuracy of the monitoring data.

[0010] Another purpose of the present invention is to solve the problem that in complex construction environments, vibration sensor data is easily interfered with by multiple factors, and it is necessary to accurately judge whether the data mutation is a real disturbance, so as to dynamically adjust the length of the sliding time window and optimize data processing.

[0011] Another object of the present invention is to solve the problem of how to determine the reasonable arrangement of vibration sensors and tilt sensors in the manual area and the surface of the support structure; Another object of the present invention is to solve the problem of how to establish a scientific method to convert soil pressure data into simulated displacement and compare it with the actual inclination, so as to timely discover the risk of instability of the support structure unit.

[0012] Another purpose of the present invention is to solve how to integrate geological radar detection data to more accurately reflect the soil conditions of the foundation pit side walls and improve the accuracy of lateral earth pressure calculations.

[0013] Another object of the present invention is to solve the problem of how to establish a mechanism for dynamically adjusting the length of the sliding time window to balance monitoring accuracy and resource consumption.

[0014] Another object of the present invention is to solve the problem of how to achieve efficient security management and emergency response through mobile terminals.

[0015] Another object of the present invention is to provide a safety management system for dangerous and major projects, which can accurately monitor the safety status of foundation pits in real time through multi-sensor collaborative monitoring, edge computing and dynamic early warning, detect anomalies and issue early warnings in a timely manner, and ensure the safety of dangerous and major project construction.

[0016] In order to achieve these purposes and other advantages according to the present invention, a dangerous and major project safety management system is provided, comprising: A plurality of vibration sensors are evenly arranged at a boundary of a construction area, wherein the construction area has only one foundation pit; A plurality of tilt sensors are evenly arranged on the surface of the support structure of the foundation pit; A plurality of static levels, which are arranged corresponding to a plurality of settlement monitoring points of the support structure, and the settlement monitoring points are evenly arranged; A plurality of earth pressure boxes are buried in layers downwardly from the side walls of the foundation pit; The regional controller is equipped with an edge computing unit to receive real-time monitoring data from vibration sensors, tilt sensors, static levels, and earth pressure cells; Among them, the edge computing unit uses its built-in finite element analysis model, takes the material elastic modulus and cross-sectional dimensions of the support structure as input parameters, and uses the layered monitoring values of each vibration sensor, tilt sensor, static level and soil pressure box as boundary conditions, and outputs the theoretical values of the vibration intensity, tilt, settlement and soil pressure data of each layer at each position in real time. When the deviation between the real-time monitoring data of each vibration sensor, tilt sensor, static level and soil pressure box of each layer and the theoretical values of the vibration intensity, tilt, settlement and soil pressure data at each position exceeds the first threshold, the current monitoring data and the mean of the historical data are fused, and the fused data is input into the finite element analysis model as the boundary condition to calculate the new theoretical values of each parameter. If, after three iterations, the deviation between each real-time monitoring design and the new theoretical value still exceeds the second threshold, a first-level warning signal is generated, 1.5 times the first threshold < the second threshold < 2 times the first threshold.

[0017] Preferably, the specific method for fusing the mean of the current monitoring data and the historical data is: Using a sliding time window, after calculating the mean of historical data according to the time decay weight, the weight coefficient of the current monitoring data and the mean of historical data is assigned: ; Among them, B is the weight coefficient of the current monitoring data, 1-B is the weight coefficient of the historical data mean, A is the current deviation, and the sliding time window length is set to 5-10 minutes for vibration sensors, 30-60 minutes for earth pressure cells, 6-12 hours for tilt sensors, and 8-16 hours for static levels.

[0018] Preferably, the safety management system for dangerous and major projects also includes a step of filtering abnormal data: if the current monitoring data exceeds the range of ±3 times the standard deviation of the historical data mean, it is determined to be abnormal data, and the abnormal data does not participate in the fusion, and the historical mean is directly used as the boundary condition to input the finite element analysis model to calculate the theoretical value.

[0019] Preferably, the hazardous engineering safety management system further includes a sensor calibration step: calculating the residual between the fused data and the theoretical value calculated thereby, and triggering automatic calibration of the corresponding sensor if the calculated residual does not decrease after two consecutive fusions.

[0020] Preferably, if the correlation between the pressure changes of three adjacent earth pressure cells is greater than 0.8 when the vibration sensor data suddenly changes, it is determined to be a real disturbance, and the length of the sensor sliding time window is shortened to the minimum value.

[0021] Preferably, any vibration sensor has a built-in triaxial accelerometer, and the distance between adjacent vibration sensors is 5-6 m; the distance between adjacent tilt sensors is 10-12 m.

[0022] Preferably, soil pressure boxes are buried every 1 to 2 meters on the side wall of the foundation pit, and the measured soil pressure values of the soil pressure boxes at each layer are recorded. The soil pressure data of each layer are weighted and fitted into a pressure gradient distribution according to the depth. The non-uniformly distributed load is used as the boundary condition to calculate the simulated displacement of each unit in the support structure. If the deviation between the simulated displacement and the actual inclination exceeds the third threshold, a secondary warning signal is generated to indicate that the support structure unit is unstable. The first threshold < the third threshold < the second threshold. The specific steps are: S1. Convert the n-layer depth Zi into the relative depth value Zi', Zi'=Zi / H; S2, assign weights Ci to each layer according to depth, ; S3. Construct a function model p(Z')=aZ'+b for the pressure variation with relative depth, and solve the parameters a and b to minimize the weighted residual square sum; S4. Discretize the surface of the support structure into a finite element grid. The position of each unit is relative to the depth Z', and the pressure value p(Z') at that location is calculated according to the function model of step S3. S5. In the finite element analysis unit, the pressure function model p(Z') is applied to the surface of the support structure as a surface force, the simulated displacement of each unit of the support structure is calculated, and compared with the real-time tilt deviation; Where H is the total depth of the foundation pit, m; Z' is the relative depth.

[0023] Preferably, in step S3, the function model of pressure variation with relative depth also incorporates geological radar detection data, specifically by: S31. Pre-embed a coaxial cable waveguide sensor in the side wall of the foundation pit to obtain the dielectric constant profile of the soil on the side wall of the foundation pit in real time; S32. Inversely calculate the soil moisture content D of the foundation pit sidewall based on the node constant, and integrate the function model of the pressure correction with the relative depth based on the soil moisture content D of the foundation pit sidewall: p(Z')=aZ'+b+eD(Z'); Where e is the correction coefficient of the lateral earth pressure due to the moisture content of the soil on the side wall of the foundation pit. The e value can be updated online using the least squares method based on the real-time monitored moisture content-earth pressure data.

[0024] Preferably, the length of the sliding time window is also dynamically adjusted according to the construction stage: Excavation stage: The vibration sensor window is shortened to 3-5 minutes, and the earth pressure cell window is shortened to 15-30 minutes; Support stage: The tilt sensor window is extended to 18~24h, and the static level window is extended to 24~48h.

[0025] Preferably, the dangerous and major project safety management system further includes: A mobile terminal receives the primary and secondary warning signals generated by the zone controller in real time. The mobile terminal marks the boundary of the secondary warning zone on the electronic plan. The boundary of the warning zone extends outward by 3 to 5 meters based on the position of the support structure unit to form a closed polygon. The mobile terminal has built-in: A security control module deploys an electronic fence in the closed polygon, the electronic fence is linked to the access controller, and when the mobile terminal receives both the first-level warning signal and the second-level warning signal, the entrance and exit gates of the warning area are closed; An emergency response module, connected to the on-site sound and light alarm, activates different alarm modes according to the warning level. A level one warning triggers a yellow rotating warning light and an intermittent buzzer, while a level two warning triggers a red strobe warning light and a continuous buzzer; Among them, the sound and light alarm is synchronously encoded with the micro linear motor embedded in the lining of the safety helmet. When the first-level warning is triggered, the micro linear motor vibrates intermittently at a frequency of 2Hz, and when the second-level warning is triggered, the micro linear motor vibrates continuously at a frequency of 5Hz.

[0026] The present invention has at least the following beneficial effects: First, the safety management system for hazardous projects provided by the present invention uses multi-type sensor collaborative monitoring, in conjunction with edge computing units and finite element analysis models, to achieve real-time, accurate monitoring and dynamic analysis of multiple parameters of foundation pits. It can promptly detect deviations between monitoring data and theoretical values, and through data fusion and iterative calculations, accurately generate a first-level early warning signal, greatly improving the accuracy and timeliness of foundation pit safety monitoring and effectively preventing accidents. Secondly, the dangerous and major engineering safety management system provided by the present invention sets the sliding time window length and weight coefficient according to the characteristics of different sensors, rationally integrates the current and historical data averages, makes data processing more in line with the actual monitoring situation, improves the reliability of monitoring data and the scientific nature of analysis results, provides a more accurate data foundation for subsequent finite element analysis based on fused data, and enhances the accuracy of the system's judgment on the safety status of foundation pits; Third, the dangerous and major project safety management system provided by the present invention also establishes an abnormal data filtering mechanism to prevent abnormal data from interfering with the monitoring results, ensure that the data input into the finite element analysis model is authentic and reliable, improve the credibility of the analysis results, and make the theoretical values calculated based on the model more reflective of the actual conditions of the foundation pit, thereby ensuring the accuracy of the system warning and effectively reducing the risk of false alarms; Fourthly, the safety management system for hazardous and major projects provided by the present invention triggers sensor calibration based on the residual difference between fused data and theoretical values. This can promptly detect potential sensor problems, ensure the long-term stable operation of the sensors, continuously output accurate monitoring data, and maintain the system's high precision in monitoring foundation pit safety. This ensures that during long-term construction, the system can always accurately reflect the safety status of the foundation pit, providing reliable protection for construction safety. Fifth, the safety management system for dangerous and major projects provided by the present invention accurately identifies real disturbances by determining the correlation between sudden changes in vibration sensor data and pressure changes in adjacent soil pressure cells. This allows for timely shortening of the sliding time window, increasing the frequency of data acquisition, and more sensitive capture of real disturbances around the foundation pit. This provides strong support for the system to quickly and accurately determine the safety status of the foundation pit, thereby improving the timeliness of early warnings. Sixth, the safety management system for dangerous and major projects provided by the present invention fits soil pressure data into a pressure gradient distribution weighted by depth, and converts it into simulated displacement and actual tilt. This can accurately assess the stability of support structure units and promptly generate secondary early warning signals to indicate the risk of instability of support structure units, buying time for construction personnel to take reinforcement measures, effectively ensuring the safety of foundation pit support structures and reducing the probability of collapse accidents. Seventh, the safety management system for hazardous and major projects provided by the present invention also incorporates a model of pressure-versus-depth function modified by geological radar detection data, fully considering the impact of soil moisture content on lateral earth pressure on the foundation pit sidewall. This makes the model more consistent with actual soil conditions, improves the accuracy of lateral earth pressure calculations, provides more precise data for support structure stability analysis, and enhances the reliability of the system for foundation pit safety monitoring and assessment. Eighth, the mobile terminal provided in the dangerous and major project safety management system provided by the present invention receives early warning signals in real time, marks the boundaries of the warning area through an electronic plane map, combines the safety management module with the emergency response module, realizes efficient safety management and graded alarm, utilizes the sound and light alarm and the helmet micro linear motor for synchronous encoding, notifies relevant personnel in multiple dimensions, effectively improves the early warning response speed and safety, and reduces the risk of accidents.

[0027] Other advantages, objectives and features of the present invention will be reflected in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a schematic diagram of the workflow of the dangerous and major project safety management system according to one technical solution of the present invention; Figure 2 This is a flow chart of the secondary warning triggering process described in another technical solution of the present invention. DETAILED DESCRIPTION

[0029] The present invention will be further described in detail below in conjunction with specific embodiments so that those skilled in the art can implement the invention with reference to the description.

[0030] It should be understood that terms such as “having”, “including” and “comprising” used herein do not exclude the existence or addition of one or more other elements or combinations thereof.

[0031] like Figure 1 As shown, the present invention provides a dangerous and major project safety management system, comprising: A plurality of vibration sensors are evenly arranged at a boundary of a construction area, wherein the construction area has only one foundation pit; A plurality of tilt sensors are evenly arranged on the surface of the support structure of the foundation pit; A plurality of static levels, which are arranged corresponding to a plurality of settlement monitoring points of the support structure, and the settlement monitoring points are evenly arranged; A plurality of earth pressure boxes are buried in layers downwardly from the side walls of the foundation pit; The regional controller is equipped with an edge computing unit to receive real-time monitoring data from vibration sensors, tilt sensors, static levels, and earth pressure cells; Among them, the edge computing unit uses its built-in finite element analysis model, takes the material elastic modulus and cross-sectional dimensions of the support structure as input parameters, and uses the layered monitoring values of each vibration sensor, tilt sensor, static level and soil pressure box as boundary conditions, and outputs the theoretical values of the vibration intensity, tilt, settlement and soil pressure data of each layer at each position in real time. When the deviation between the real-time monitoring data of each vibration sensor, tilt sensor, static level and soil pressure box of each layer and the theoretical values of the vibration intensity, tilt, settlement and soil pressure data at each position exceeds the first threshold, the current monitoring data and the mean of the historical data are fused, and the fused data is input into the finite element analysis model as the boundary condition to calculate the new theoretical values of each parameter. If, after three iterations, the deviation between each real-time monitoring design and the new theoretical value still exceeds the second threshold, a first-level warning signal is generated, 1.5 times the first threshold < the second threshold < 2 times the first threshold.

[0032] In the above implementation scheme, the sensors are arranged as follows: The vibration sensor uses a high-precision MEMS vibration sensor, one optional model being the Bosch BMI160. Multiple vibration sensors of this type are evenly arranged at the boundaries of the construction area at a certain interval (preferably 5-6m), and the collected vibration data is transmitted in real time to the regional controller via a wireless transmission module. The tilt sensor uses a capacitive tilt sensor, one optional model being the SCA100T, with an accuracy of up to ±0.01°. These tilt sensors are evenly installed at certain intervals (preferably 10-12m) on the surface of the foundation pit support structure, and the tilt data is also transmitted to the regional controller via a wireless transmission module. The static level uses a high-precision model, one optional model being the DS05, with a measurement accuracy of up to ±0.5mm. It is installed at settlement monitoring points evenly distributed on the support structure and connected to the regional controller via an RS485 bus to ensure the stability and accuracy of data transmission. The soil pressure box uses a high-precision soil pressure box. One optional model is the ZP-100, which has a measuring range of 0~5MPa and an accuracy of up to ±0.5%FS. The soil pressure boxes are buried in layers on the side wall of the foundation pit and connected to the regional controller through cables to transmit the soil pressure data of each layer.

[0033] In the above-mentioned implementation scheme, the regional controller utilizes a high-performance industrial-grade embedded computer, one optional model being Advantech's ARK-3500. Its built-in edge computing unit utilizes an Intel Core i7 processor, offering powerful computing capabilities. By programming a dedicated driver, it enables real-time reception of sensor data. The finite element analysis model within the edge computing unit is constructed using the open-source FEniCS software framework. Parameters such as the material elastic modulus and cross-sectional dimensions of the support structure are input into the model. In combination with the real-time monitoring values collected by each sensor as boundary conditions, optimized numerical algorithms (such as the conjugate gradient method) are used to calculate and output in real time the theoretical values of vibration intensity, tilt, settlement, and soil pressure data for each layer at each location.

[0034] In the above implementation, the monitoring data from each sensor is compared in real time with the theoretical value output by the finite element analysis model. When the deviation between the two exceeds a first threshold, the data fusion mechanism is activated. Historical data from a certain period of time is retrieved from the historical data storage database, and the mean of the historical data is calculated. This data is then fused with the current monitoring data according to a certain weight. The fused data is then input into the finite element analysis model as a boundary condition to calculate a new theoretical value. If, after three iterations, the deviation between the real-time monitoring data and the new theoretical value still exceeds a second threshold, the regional controller sends a level 1 warning signal to the designated receiving terminal via limited or wireless communication, where 1.5 times the first threshold < the second threshold < 2 times the first threshold.

[0035] In the above-mentioned embodiment, the MEMS vibration sensor may employ a fiber Bragg grating (FBG) vibration sensor. Fiber Bragg grating (FBG) vibration sensors have the advantages of strong resistance to electromagnetic interference and high precision, and can operate more stably in complex electromagnetic environments. They sense vibrations by measuring wavelength changes in the fiber Bragg grating (FBG). Wavelength division multiplexing (WDM) technology can be used to connect multiple sensors in series on a single optical fiber, reducing wiring complexity but at a higher cost. The embodiments of the present invention may also utilize a cloud computing platform in place of an edge computing unit for data processing and finite element analysis. This platform has more powerful computing capabilities, but may suffer from data transmission delays and network stability issues, necessitating network coverage.

[0036] In the above implementation scheme, the present invention uses a variety of sensors to monitor the foundation pit from different dimensions. Vibration sensors monitor the vibration conditions around the construction area, tilt sensors monitor the tilt of the support structure, static levels monitor settlement, and soil pressure boxes monitor soil pressure, achieving comprehensive coverage of the safety status of the foundation pit. Data on all aspects of the foundation pit can be accurately obtained, providing a rich and accurate data basis for subsequent analysis. At the same time, an edge computing unit is used in combination with real-time processing of sensor data to dynamically calculate theoretical values based on actual monitoring values. Compared with the traditional method of periodic manual data collection and subsequent analysis, the timeliness of the analysis is greatly improved, and the changing trend of the foundation pit safety status can be discovered in a timely manner. Through data deviation processing and iterative calculation, an early warning is triggered only when the deviation persists and exceeds a certain threshold, effectively avoiding false alarms caused by accidental interference factors, improving the accuracy and reliability of the early warning, and providing a strong guarantee for the safe construction of dangerous and major projects.

[0037] According to the above embodiments, the present invention has at least the following beneficial effects: 1. This implementation plan uses multiple vibration sensors evenly distributed at the construction area boundary, tilt sensors on the support structure surface, static levels at corresponding settlement monitoring points, and earth pressure cells buried in layers on the pit sidewalls to comprehensively monitor the foundation pit from multiple dimensions: vibration, tilt, settlement, and earth pressure. This accurately captures state change data for all parts and aspects of the pit, providing a rich and accurate data foundation for subsequent safety assessments. This significantly improves the comprehensiveness and accuracy of monitoring, preventing safety hazards from being overlooked due to blind spots or missing data. 2. In this implementation, the edge computing unit within the domain controller receives sensor data in real time and uses a built-in finite element analysis model, combined with support structure parameters and real-time monitoring values, to rapidly calculate the theoretical values of relevant data at each location. This real-time data processing and analysis significantly shortens the data processing cycle compared to traditional manual data collection and analysis, enabling timely detection of dynamic changes in the pit's safety status and buying valuable time for timely response measures. 3. In this implementation plan, when the deviation between the real-time monitoring data and the theoretical value exceeds the first threshold, the data fusion mechanism is activated to fuse the current monitoring data with the average of historical data and recalculate the theoretical value. If the deviation still exceeds the second threshold after three iterations, a warning signal is issued. This effectively eliminates the interference of accidental factors, avoids frequent false alarms, and ensures the accuracy and reliability of the warning. The scientific and reasonable warning mechanism enables relevant personnel to accurately judge the actual safety status of the foundation pit, so as to take corresponding safety measures in a timely and accurate manner to ensure the safety of dangerous and major project construction.

[0038] In one embodiment, the specific method of fusing the mean of the current monitoring data and the historical data is: Using a sliding time window, after calculating the mean of historical data according to the time decay weight, the weight coefficient of the current monitoring data and the mean of historical data is assigned: ; Among them, B is the weight coefficient of the current monitoring data, 1-B is the weight coefficient of the historical data mean, A is the current deviation, and the sliding time window length is set to 5-10 minutes for vibration sensors, 30-60 minutes for earth pressure cells, 6-12 hours for tilt sensors, and 8-16 hours for static levels.

[0039] In the above implementation scheme, vibration sensors, tilt sensors, static and earth pressure boxes continuously collect data and transmit it to the regional controller in real time, and store it in a local time series database. The database stores data in a structured manner according to dimensions such as sensor type and timestamp, providing a data basis for calculating the mean of historical data and performing data fusion.

[0040] In the above embodiments, for the vibration sensor, a sliding time window of 5 to 10 minutes is set. Within the window, the mean value of historical data is calculated according to the time-decaying weight. The data closer to the current time has a higher weight, and the data farther from the current time has a lower weight. Taking 8 minutes as an example, the weight of the data in the most recent 1 minute is 0.5, the weight of the data in the 2nd minute is 0.3, the weight of the data in the 3rd minute is 0.15, and the total weight of the data from the 4th to the 8th minute is 0.05. For the earth pressure cell, a sliding time window of 30 to 60 minutes is set. Since the change of earth pressure is relatively slow, a longer time window can better reflect its long-term trend. Within the time window, the mean value of historical data is also calculated according to the time-decaying weight. Taking 50 minutes as an example, the total weight of the data in the first 10 minutes is 0.2, the total weight of the data in the middle 20 minutes is 0.5, and the total weight of the data in the last 20 minutes is 0.3, and the weights of the data in each internal time period are distributed according to exponential decay. For the tilt sensor, a sliding time window of 6 to 12 hours is set. Because the tilt change of the support structure is usually relatively slow, a long window can more accurately reflect its change trend. Within the time window, the mean value of historical data is also calculated according to the time-decaying weight. Taking a 9-hour window as an example, the total weight of the data in the first 3 hours is 0.2, the total weight of the data in the middle 3 hours is 0.5, and the total weight of the data in the last 3 hours is 0.3, and the weights of the data within each hour are further subdivided according to exponential decay. For the static level, a sliding time window of 8 to 16 hours is set. Within the time window, because the change of settlement data is slow, a long window is beneficial to smooth the data fluctuation and accurately obtain the settlement trend. The mean value of historical data is also calculated according to the time-decaying weight. Taking a 12-hour window as an example, the total weight of the data in the first 4 hours is 0.2, the total weight of the data in the middle 4 hours is 0.5, and the total weight of the data in the last 4 hours is 0.3, and the weights of the data within each hour also decay exponentially.

[0041] In the above embodiments, after calculating the deviation between the current monitoring data and the theoretical value, the weight coefficients of the current monitoring data and the mean value of historical data are allocated according to the size of the deviation. If A ≤ 5%, the weight coefficient B of the current monitoring data is set to 0.7, and the weight coefficient 1 - B of the mean value of historical data is 0.3. This means that when the deviation is small, the system trusts the current monitoring data more. When 5% < A < 10%, the weight coefficient is calculated by the formula B = 0.7 - 0.4(A - 5%). If A = 7%, then B = 0.7 - 0.4×(7% - 5%) = 0.692, and 1 - B = 0.308. As the deviation increases, the weight of the current monitoring data gradually decreases, and the weight of the mean value of historical data increases correspondingly to balance the reliability of data fusion.

[0042] In the above implementation, the current monitoring data is further fused with the mean historical data based on the calculated weight coefficients. Taking a vibration sensor as an example, the current 8-minute monitoring data, weighted B, is weighted and summed with the mean historical data calculated within the sliding time window using the time-decayed weights, weighted 1-B, to produce the fused data. This fused data is then input into the finite element analysis model as boundary conditions, and the new theoretical values of each parameter are recalculated for comparison with the real-time monitoring data to determine whether further iteration or the triggering of an early warning is necessary.

[0043] In the above implementation scheme, the sliding time window can be adjusted based on fuzzy logic, which can more flexibly adjust the sliding time window according to the real-time characteristics of the data, adaptively optimize the data fusion process in different construction stages and different environmental conditions, and improve the ability to respond to complex situations. The definition of three fuzzy variables and the formulation of fuzzy reasoning rules require certain professional knowledge and experience, and need to be continuously optimized and adjusted to adapt to different construction scenarios.

[0044] In the above implementation, sliding time windows of varying lengths are set to address the rapid changes in vibration sensor data, the relatively slow changes in earth pressure cell data, and the even slower changes in tilt sensor and static level data. This allows for more efficient utilization of historical and current data, accurately reflecting the changing trends of various parameters. For example, the short time window of the vibration sensor can promptly capture rapidly changing vibration conditions, while the long time window of the earth pressure cell can smooth out its slowly changing data, improving the accuracy of data fusion. Furthermore, the weighting coefficients of the current monitoring data and the mean of historical data are dynamically adjusted based on the magnitude of the deviation, prioritizing the current data when the deviation is small and increasing the weight of the mean of historical data when the deviation increases. This dynamic adjustment mechanism enables the system to better cope with data fluctuations and anomalies, improving the reliability and stability of data fusion. For example, when the deviation is small, the current data can more promptly reflect the current status; when the deviation is large, the mean of historical data can assist in determining whether the anomaly is a true one, avoiding misjudgments. Finally, through appropriate data fusion and calculation of new theoretical values, the safety status of the foundation pit can be more accurately determined, reducing false alarms caused by data fluctuations or accidental factors. During the iterative calculation process, the theoretical value is continuously optimized, resulting in more accurate warnings and providing reliable protection for the construction safety of hazardous and major projects. For example, by calculating theoretical values through multiple iterations of data fusion, safety hazards in foundation pits can be identified more accurately, avoiding interference with construction caused by false alarms.

[0045] In one of the implementation schemes, the safety management system for dangerous and major projects also includes a step of filtering abnormal data: if the current monitoring data exceeds the range of ±3 times the standard deviation of the historical data mean, it is judged as abnormal data. The abnormal data does not participate in the fusion, and the historical mean is directly used as the boundary condition to input the finite element analysis model to calculate the theoretical value.

[0046] In the above-described embodiment, the regional controller's database stores data in time series, simultaneously recording information such as each sensor's unique identifier, measurement value, and measurement time. The regional controller regularly calculates the mean and standard deviation of each sensor's historical data. When new monitoring data is generated, it is immediately compared with the corresponding historical data mean and standard deviation. If the historical data mean - 3 times the standard deviation < the current monitoring data < the historical data mean + 3 times the standard deviation, the data is considered abnormal. Once identified as abnormal, the data is not included in the fusion process of the current monitoring data and the historical data mean. Instead, the historical mean is directly used as the boundary condition for input into the finite element analysis model to calculate the theoretical value.

[0047] In the above implementation plan, by filtering out abnormal data that exceeds the range of ±3 times the standard deviation of the historical data mean, the influence of abnormal values caused by sensor failure, interference, etc. on data fusion and theoretical value calculation is avoided, making the data input into the finite element analysis model more reliable, thereby improving the accuracy of the assessment of the safety status of the foundation pit. Directly using the historical mean as the input in the case of abnormal data maintains the relative stability of the input data of the finite element analysis model. The occasional abnormal data will not cause the model calculation results to fluctuate significantly, ensuring that the system can still operate stably when abnormal data appears, and continuously and accurately assess the safety status of the foundation pit. Normal data may often cause the system to falsely report foundation pit safety issues. By filtering abnormal data, false warnings caused by abnormal values are reduced, so that the warning signal can more truly reflect the actual safety status of the foundation pit, avoiding unnecessary interference to construction due to frequent false alarms.

[0048] In one embodiment, the hazardous engineering safety management system further includes a sensor calibration step: calculating the residual between the fused data and the theoretical value calculated thereby; if the calculated residual does not decrease after two consecutive fusions, the automatic calibration of the corresponding sensor is triggered.

[0049] In the above implementation scheme, after the current monitoring data is fused with the mean of the historical data, the fused data is input as boundary conditions into the finite element analysis model, and the theoretical values of the parameters at each location are calculated. The residuals of the fused data and the theoretical values calculated from the fused data are further calculated and stored in the database of the regional controller. The residuals are associated with the corresponding fused data, theoretical values, timestamps, and other information. If the residuals obtained from two consecutive fusion calculations do not decrease, the automatic calibration process of the corresponding sensor is triggered. For vibration sensors, a standard vibration source can be used for calibration. Tilt sensors can be calibrated by comparing them with a high-precision horizontal reference. Static levels can be calibrated using a standard liquid level gauge. Earth pressure cells can be calibrated by loading standard weights with known pressures.

[0050] In this implementation, by continuously monitoring the residuals between the fused data and the theoretical values and performing sensor calibration when the residuals continue to decline, potential sensor measurement deviations can be promptly detected, ensuring that the sensors consistently output accurate and reliable data. Accurate monitoring data is crucial for assessing the safety status of foundation pits based on finite element analysis models, effectively avoiding misjudgments of safety hazards caused by sensor errors. Furthermore, the automatic calibration mechanism maintains the stability of the system's monitoring data, preventing deviations in the system's assessment of the pit's safety status due to sensor performance drift. A stable and reliable monitoring system can consistently and accurately reflect the actual safety status of the foundation pit, providing construction personnel with a reliable basis for decision-making and ensuring the smooth progress of critical construction projects. Compared to traditional periodic manual sensor calibration methods, this automatic calibration mechanism intelligently triggers calibration based on the residuals of actual monitoring data, eliminating the need for frequent manual inspection and calibration of sensors and significantly reducing manual intervention costs. Furthermore, automatic calibration enables timely response to issues as they arise, avoiding inaccurate monitoring data caused by untimely manual calibration, thereby improving system efficiency.

[0051] In one embodiment, if the correlation between the pressure changes of three adjacent earth pressure cells is greater than 0.8 when the vibration sensor data suddenly changes, it is determined to be a real disturbance, and the length of the sensor sliding time window is shortened to the minimum value.

[0052] In the above-mentioned implementation scheme, the vibration sensors continuously monitor vibration data at the construction zone boundary, while the earth pressure cells monitor earth pressure data at each layer of the foundation pit sidewall in real time. The regional controller receives and stores this data in real time, recording the vibration intensity values of the vibration sensors and the pressure values of the earth pressure cells in a time series format, and also noting information such as the acquisition time and sensor number for each data point. The regional controller analyzes the vibration sensor data in real time, calculating the rate of change of each vibration sensor data point. If the rate of change of vibration intensity for a particular vibration sensor exceeds a preset mutation threshold (generally set to a certain multiple of the normal fluctuation range, such as the historical mean plus or minus a certain number of standard deviations), the vibration sensor data point is determined to have experienced a mutation. Upon detecting a mutation in vibration sensor data, the system immediately extracts pressure data from the three adjacent earth pressure cells for the same time period. Using a correlation coefficient calculation method, such as the Pearson correlation coefficient formula, the system calculates the correlation between the pressure changes of these three earth pressure cells and the mutation in the vibration sensor data point. If the calculated correlation is greater than 0.8, the disturbance is determined to be a real disturbance. If this is determined to be a real disturbance, the sliding time window length of the vibration sensor is shortened to the minimum value, based on the previously specified sliding time window length for each sensor.

[0053] In the above implementation scheme, by judging the correlation between the pressure changes of adjacent soil pressure boxes when the vibration sensor data suddenly changes, it is possible to effectively distinguish between real disturbances and false data mutations caused by sensor noise or other interference factors. When the correlation is greater than 0.8, it is determined to be a real disturbance, which improves the accuracy of identifying the real vibration conditions around the foundation pit and avoids unnecessary data processing and early warning due to misjudgment of false mutations. Once it is determined to be a real disturbance, the sliding time window length of the vibration sensor is shortened to the minimum value, so that the system can collect and analyze vibration data more frequently, which helps to capture the rapid change trend of the vibration conditions around the foundation pit in a timely manner, and provide more real-time and accurate data support for the foundation pit safety status assessment based on the finite element analysis model. The sliding time window length is shortened only when it is determined to be a real disturbance, avoiding unnecessary shortening of the window length under normal circumstances, thereby reducing the frequency and amount of calculation of data processing. While ensuring the ability to accurately monitor real disturbances, the system's computing resources are reasonably allocated, the efficiency of data processing is improved, and the operating cost of the system is reduced.

[0054] In one implementation, each vibration sensor is equipped with a built-in triaxial accelerometer, with the spacing between adjacent vibration sensors being 5-6 meters; the spacing between adjacent tilt sensors is 10-12 meters. For the vibration sensors, each built-in triaxial accelerometer simultaneously monitors vibration along three mutually perpendicular axes, enabling comprehensive and accurate acquisition of vibration information at the construction zone boundary. Compared to single-axis or dual-axis accelerometers, triaxial accelerometers can more accurately detect changes in vibration direction and intensity, providing richer and more reliable data for analyzing the impact of perimeter vibration on the support structure. The spacing between adjacent vibration sensors is set at 5-6 meters. This ensures comprehensive coverage of vibration monitoring at the construction zone boundary, avoiding blind spots, while also avoiding excessive density that wastes resources. This appropriate spacing allows the system to promptly capture vibration changes at different locations, improving the timeliness and accuracy of vibration monitoring around the foundation pit and helping to promptly identify potential vibration safety hazards. For the tilt sensors, the characteristics of the foundation pit support structure and the nature of tilt changes are fully considered. At this spacing, the tilt sensor can effectively monitor the tilt of the support structure surface and accurately reflect the deformation state of the support structure at different positions, which not only ensures the effective monitoring of the overall tilt of the support structure, but also reasonably controls the number and cost of sensors. By evenly arranging the tilt sensors, the system can have a more comprehensive understanding of the tilt distribution of the support structure and promptly detect local tilt anomalies, providing a strong data basis for evaluating the stability of the foundation pit support structure and helping to take measures in advance to prevent safety accidents such as instability of the support structure.

[0055] like Figure 2In one of the implementation schemes, soil pressure boxes are buried every 1 to 2 meters on the side wall of the foundation pit, and the measured soil pressure values of the soil pressure boxes at each layer are recorded. The soil pressure data of each layer are weighted and fitted into a pressure gradient distribution according to the depth. The non-uniformly distributed load is used as the boundary condition to calculate the simulated displacement of each unit in the support structure. If the deviation between the simulated displacement and the actual inclination exceeds the third threshold, a secondary warning signal is generated to indicate that the support structure unit is unstable. The first threshold < the third threshold < the second threshold. The specific steps are as follows: S1. Convert the n-layer depth Zi into the relative depth value Zi', Zi'=Zi / H; S2, assign weights Ci to each layer according to depth, ; S3. Construct a function model p(Z')=aZ'+b for the pressure variation with relative depth, and solve the parameters a and b to minimize the weighted residual square sum; S4. Discretize the surface of the support structure into a finite element grid. The position of each unit is relative to the depth Z', and the pressure value p(Z') at that location is calculated according to the function model of step S3. S5. In the finite element analysis unit, the pressure function model p(Z') is applied to the surface of the support structure as a surface force, the simulated displacement of each unit of the support structure is calculated, and compared with the real-time tilt deviation; Where H is the total depth of the foundation pit, m; Z' is the relative depth.

[0056] In the above implementation scheme, first, the measured values of the soil pressure of each layer are recorded in real time according to the soil pressure boxes buried in the side walls of the foundation pit, and at the same time, the support surface tilt sensor records the actual tilt data in real time. Then, the depth of the layer where the soil pressure boxes are located is converted into a relative depth value, and the depth data of different foundation pits are normalized to facilitate subsequent calculations and analysis. Further, each layer is given a corresponding weight according to its relative depth. Taking into account the different degrees of influence of soil layers at different depths on the support structure, the weighted method is used to more accurately reflect the comprehensive effect of the soil pressure of each layer on the support structure. Then, a function model of the pressure change with relative depth is constructed, and a linear function is used to approximate the change law of soil pressure with depth. Then, by solving the parameters a and b, the sum of squares of the weighted residuals is minimized. The surface of the support structure is then discretized into a finite element grid, each unit has its corresponding relative depth, and the pressure value at each unit position is calculated according to the above function model. In the finite element analysis unit, the pressure function model is applied to the surface of the support structure as a surface force, and the simulated displacement of each unit of the support structure is calculated using the finite element analysis model. Finally, the calculated simulated displacement is compared with the real-time tilt. If the deviation exceeds the third threshold (first threshold < third threshold < second threshold), a secondary warning signal is generated, indicating that the support structure unit is unstable.

[0057] In this implementation, depth-weighted fitting of soil pressure data for each layer takes into account the varying effects of soil layers at different depths on the support structure. This allows for more accurate simulation of the actual distribution of soil pressure, thereby improving the accuracy of finite element analysis. By comparing simulated displacements with actual tilt and setting appropriate thresholds, potential instability of support structure units can be promptly detected, issuing early warning signals and providing a crucial basis for project safety management. Converting depth data to relative depth values makes analysis results for different foundation pits comparable, facilitating unified safety assessment and management across different projects.

[0058] In one embodiment, in step S3, the function model of pressure variation with relative depth is further integrated with geological radar detection data, specifically in the following manner: S31. Pre-embed a coaxial cable waveguide sensor in the side wall of the foundation pit to obtain the dielectric constant profile of the soil on the side wall of the foundation pit in real time; S32. Inversely calculate the soil moisture content D of the foundation pit sidewall based on the node constant, and integrate the function model of the pressure correction with the relative depth based on the soil moisture content D of the foundation pit sidewall: p(Z')=aZ'+b+eD(Z'); Where e is the correction coefficient of the lateral earth pressure due to the moisture content of the soil on the side wall of the foundation pit. The e value can be updated online using the least squares method based on the real-time monitored moisture content-earth pressure data.

[0059] In the above-mentioned implementation scheme, a coaxial cable waveguide sensor is embedded in the pit sidewall. This sensor can acquire the dielectric constant profile of the soil in the pit sidewall in real time. The coaxial cable waveguide sensor transmits electromagnetic waves into the soil and receives the reflected signal. Based on the changes in the signal, the distribution of the soil dielectric constant is determined and transmitted to the regional controller for storage and subsequent processing. The acquired soil dielectric constant data is inverted based on the node constants to obtain the moisture content D of the pit sidewall soil. The specific steps of this step are: using an empirical formula (such as the Topp equation) or a physical model (such as the Debye model) to establish a mapping between dielectric constant and moisture content. The dielectric constant data obtained by the coaxial cable waveguide sensor is first filtered to reduce noise. Then, the temperature drift of the dielectric constant is corrected based on the data from the sensor's built-in temperature module. The model coefficients (such as the coefficients in the Topp equation or the parameters in the physical model) are adjusted using the least squares method. The calibrated model is then embedded in the regional controller to output moisture content data in real time. Finally, the calculated soil moisture content D of the foundation pit side wall is integrated into the function model of pressure changing with depth.

[0060] In the above implementation, geological radar data is integrated into the model that modifies the pressure-with-depth function, taking into account the effect of soil moisture content on lateral earth pressure. Changes in soil moisture content significantly affect the magnitude of earth pressure. By acquiring real-time moisture content data and incorporating it into the model, earth pressure can be calculated more accurately, making finite element analysis results more realistic and improving the accuracy of the mechanical performance assessment of the support structure. By using an online update of the correction factor e, the pressure function model can be adjusted in real time based on the dynamic changes in soil moisture content during actual construction. This enables the system to adapt to different geological conditions and construction conditions, improving its adaptability and reliability and enabling more effective monitoring of the stability of the support structure. More accurate earth pressure calculations and more realistic finite element analysis results provide a more reliable basis for determining whether support structure units are unstable. By generating timely and accurate secondary warning signals, potential safety hazards can be detected in advance, buying more time to implement appropriate safety measures and ensuring the construction safety of critical and major projects.

[0061] In one embodiment, the length of the sliding time window is also dynamically adjusted according to the construction stage: Excavation stage: The vibration sensor window is shortened to 3-5 minutes, and the earth pressure cell window is shortened to 15-30 minutes; Support phase: The tilt sensor window is extended to 18~24 hours, and the static level window is extended to 24~48 hours. In the above implementation scheme, during the excavation phase, operations such as earthwork excavation and blasting frequently induce vibrations, which may impact the surrounding environment and support structures of the foundation pit. Shortening the vibration sensor's sliding time window from the conventional 5-10 minutes to 3-5 minutes allows for more timely detection of subtle changes in vibration. During excavation, the stress state of the soil on the pit sidewalls continuously changes, and so does the soil pressure. Shortening the sliding time window of the soil pressure cell from 30-60 minutes to 15-30 minutes allows for more real-time monitoring of dynamic changes in soil pressure. Promptly detecting abnormal increases or decreases in soil pressure helps to identify signs of soil instability, such as landslides or collapses, allowing for timely adjustments to construction plans and avoiding safety incidents. The support phase primarily involves the construction and reinforcement of the foundation pit's support structures, and their deformation is a relatively slow process. Extending the tilt sensor's sliding time window from 6-12 hours to 18-24 hours allows for more comprehensive observation of the tilt trends of the support structure over a longer period of time. Avoid misjudgments due to small fluctuations in a short period of time and more accurately assess the stability of the support structure. The static level is used to monitor the settlement of the support structure. During the support stage, the development of settlement is also a relatively slow process. Extending the sliding time window of the static level from 8~16h to 24~48h can better capture the long-term trend of settlement. For some subtle settlement changes, long-term monitoring can more accurately analyze their development speed and stability, providing a reliable basis for judging the safety of the support structure. If the settlement shows a stable and slow growth trend over a long period of time, it means that the support structure is in a normal state; if the settlement suddenly accelerates or exceeds a certain threshold, an early warning can be issued in time and appropriate treatment measures can be taken.

[0062] In the above implementation plan, by dynamically adjusting the length of the sliding time window according to the construction stage, the system can perform accurate monitoring according to the actual situation at different construction stages, which not only improves the timeliness and accuracy of monitoring, but also avoids the large amount of redundant data generated by unnecessary frequent monitoring, improves the operating efficiency and reliability of the system, and thus better ensures the construction safety of dangerous projects.

[0063] In one embodiment, the hazardous project safety management system further includes: A mobile terminal receives the primary and secondary warning signals generated by the zone controller in real time. The mobile terminal marks the boundary of the secondary warning zone on the electronic plan. The boundary of the warning zone extends outward by 3 to 5 meters based on the position of the support structure unit to form a closed polygon. The mobile terminal has built-in: A security control module deploys an electronic fence in the closed polygon, the electronic fence is linked to the access controller, and when the mobile terminal receives both the first-level warning signal and the second-level warning signal, the entrance and exit gates of the warning area are closed; An emergency response module, connected to the on-site sound and light alarm, activates different alarm modes according to the warning level. A level one warning triggers a yellow rotating warning light and an intermittent buzzer, while a level two warning triggers a red strobe warning light and a continuous buzzer; Among them, the sound and light alarm is synchronously encoded with the micro linear motor embedded in the lining of the safety helmet. When the first-level warning is triggered, the micro linear motor vibrates intermittently at a frequency of 2Hz, and when the second-level warning is triggered, the micro linear motor vibrates continuously at a frequency of 5Hz.

[0064] In the above-described embodiment, the mobile terminal can receive the first- and second-level warning signals generated by the zone controller in real time and mark the boundaries of the second-level warning zone on the electronic plan. The warning zone boundary is a closed polygon formed by extending 3-5 meters outward from the position of the support structure unit. This intuitive visual display allows relevant personnel to quickly and accurately grasp the scope and location of the danger zone. Construction personnel, management personnel, and others can use the mobile terminal to quickly understand the on-site situation, plan their routes in advance, and avoid entering the danger zone, thereby effectively reducing the possibility of safety accidents. The security control module deploys an electronic fence at the closed polygon and interacts with the access control controller. When the mobile terminal receives both the first- and second-level warning signals, it automatically closes the entrance and exit gates of the warning zone. This function effectively prevents unauthorized personnel from entering the danger zone, preventing people from unknowingly falling into danger, and further ensuring the safety of construction workers. It also helps to limit the flow of people in the danger zone, facilitating subsequent emergency response and rescue efforts. Furthermore, the present invention also implements corresponding control measures based on different warning levels, making safety management more scientific and reasonable. During a Level 1 alert, although the danger level is relatively low, electronic fencing and access control systems can still alert personnel to safety. During a Level 2 alert, the timely closure of entrance and exit gates quickly cuts off access to the danger zone, minimizing casualties and property damage. The emergency response module connects to on-site audible and visual alarms and activates different alarm modes based on the alert level. A Level 1 alert triggers a yellow rotating warning light and intermittent buzzing, while a Level 2 alert triggers a red strobing warning light and continuous buzzing. This diverse alarm mode effectively draws attention in diverse environments, allowing construction workers to detect danger promptly. The yellow rotating warning light and intermittent buzzing provide a gentle reminder during a Level 1 alert, while the red strobing warning light and continuous buzzing strongly alert workers to the severity of the danger during a Level 2 alert. The audible and visual alarms are synchronized with a micro linear motor embedded in the helmet's lining. When a Level 1 alert is triggered, the micro linear motor vibrates intermittently at a 2Hz frequency, while when a Level 2 alert is triggered, the micro linear motor vibrates continuously at a 5Hz frequency. This design addresses the noisy construction site environment, where audible and visual alarms may be overlooked. Through the vibration of the helmet lining, early warning information can be directly transmitted to construction workers, ensuring that they can receive danger signals in time even in noisy environments, thereby improving the reliability of emergency response.

[0065] This implementation plan has established a comprehensive, multi-layered emergency management system, integrating and collaborating across all aspects, from the visual display of early warning information to the automated operation of safety management and control, to the diverse means of emergency response. This allows the system to rapidly respond to safety hazards during the construction of hazardous and major projects, promptly conveying hazard information to relevant personnel and implementing effective control and emergency response measures. This significantly improves the efficiency of overall emergency management and provides a solid foundation for the safe construction of hazardous and major projects.

[0066] The number of devices and processing scales described here are used to simplify the description of the present invention. Applications, modifications and variations of the hazardous engineering safety management system of the present invention will be obvious to those skilled in the art.

[0067] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the embodiments shown and described herein.

Claims

1. Dangerous and major project safety management system, characterized by: include: A plurality of vibration sensors are evenly arranged at a boundary of a construction area, wherein the construction area has only one foundation pit; A plurality of tilt sensors are evenly arranged on the surface of the support structure of the foundation pit; A plurality of static levels, which are arranged corresponding to a plurality of settlement monitoring points of the support structure, and the settlement monitoring points are evenly arranged; A plurality of earth pressure boxes are buried in layers downwardly from the side walls of the foundation pit; The regional controller is equipped with an edge computing unit to receive real-time monitoring data from vibration sensors, tilt sensors, static levels, and earth pressure cells; Among them, the edge computing unit uses its built-in finite element analysis model, takes the material elastic modulus and cross-sectional dimensions of the support structure as input parameters, and uses the layered monitoring values of each vibration sensor, tilt sensor, static level and soil pressure box as boundary conditions, and outputs the theoretical values of the vibration intensity, tilt, settlement and soil pressure data of each layer at each position in real time. When the deviation between the real-time monitoring data of each vibration sensor, tilt sensor, static level and soil pressure box of each layer and the theoretical values of the vibration intensity, tilt, settlement and soil pressure data at each position exceeds the first threshold, the current monitoring data and the mean of the historical data are fused, and the fused data is input into the finite element analysis model as the boundary condition to calculate the new theoretical values of each parameter. If, after three iterations, the deviation between each real-time monitoring design and the new theoretical value still exceeds the second threshold, a first-level warning signal is generated, 1.5 times the first threshold < the second threshold < 2 times the first threshold.

2. The dangerous and major project safety management system according to claim 1 is characterized in that: The specific method of merging the mean of the current monitoring data and the historical data is as follows: Using a sliding time window, after calculating the mean of historical data according to the time decay weight, the weight coefficient of the current monitoring data and the mean of historical data is assigned: B= ； Among them, B is the weight coefficient of the current monitoring data, 1-B is the weight coefficient of the historical data mean, A is the current deviation, and the sliding time window length is set to 5-10 minutes for vibration sensors, 30-60 minutes for earth pressure cells, 6-12 hours for tilt sensors, and 8-16 hours for static levels.

3. The dangerous and major project safety management system according to claim 2 is characterized in that: It also includes the step of filtering abnormal data: if the current monitoring data exceeds the range of ±3 times the standard deviation of the historical data mean, it is judged as abnormal data. The abnormal data does not participate in the fusion, and the historical mean is directly used as the boundary condition to input the finite element analysis model to calculate the theoretical value.

4. The dangerous and major project safety management system according to claim 3 is characterized in that: It also includes a sensor calibration step: calculating the residual between the fused data and the theoretical value calculated by it. If the calculated residual does not decrease after two consecutive fusions, the automatic calibration of the corresponding sensor is triggered.

5. The dangerous and major project safety management system according to claim 4 is characterized in that: If the correlation between the pressure changes of three adjacent earth pressure cells is greater than 0.8 when the vibration sensor data suddenly changes, it is determined to be a real disturbance, and the length of the sensor sliding time window is shortened to the minimum value.

6. The dangerous and major project safety management system according to claim 1 is characterized in that: Each vibration sensor has a built-in triaxial accelerometer. The distance between adjacent vibration sensors is 5 to 6 meters; the distance between adjacent tilt sensors is 10 to 12 meters.

7. The dangerous and major project safety management system according to claim 1 is characterized in that: Earth pressure cells are buried every 1-2 meters on the sidewall of the foundation pit, and the measured earth pressure values of each layer of earth pressure cells are recorded. The earth pressure data of each layer are weighted according to the depth to fit into a pressure gradient distribution. The non-uniform load form is used as the boundary condition to calculate the simulated displacement of each unit in the support structure. If the deviation between the simulated displacement and the actual inclination exceeds the third threshold, a secondary warning signal is generated to indicate that the support structure unit is unstable. The first threshold < the third threshold < the second threshold. The specific steps are as follows: S1. Convert the n-layer depth Zi into the relative depth value Zi', Zi'=Zi / H; S2, assign weights Ci to each layer according to depth, ; S3. Construct a function model p(Z')=aZ'+b for the pressure variation with relative depth, and solve the parameters a and b to minimize the weighted residual square sum; S4. Discretize the surface of the support structure into a finite element grid. The position of each unit is relative to the depth Z', and the pressure value p(Z') at that location is calculated according to the function model of step S3. S5. In the finite element analysis unit, the pressure function model p(Z') is applied to the surface of the support structure as a surface force, the simulated displacement of each unit of the support structure is calculated, and compared with the real-time tilt deviation; Where H is the total depth of the foundation pit, m; Z' is the relative depth.

8. The dangerous and major project safety management system according to claim 7 is characterized in that: In step S3, the function model of pressure variation with relative depth is also integrated with geological radar detection data, and the specific method is as follows: S31. Pre-embed a coaxial cable waveguide sensor in the side wall of the foundation pit to obtain the dielectric constant profile of the soil on the side wall of the foundation pit in real time; S32. Inversely calculate the soil moisture content D of the foundation pit sidewall based on the node constant, and integrate the function model of the pressure correction with the relative depth based on the soil moisture content D of the foundation pit sidewall: p(Z')=aZ'+b+eD(Z'); Where e is the correction coefficient of the lateral earth pressure due to the moisture content of the soil on the side wall of the foundation pit. The e value can be updated online using the least squares method based on the real-time monitored moisture content-earth pressure data.

9. The dangerous and major project safety management system according to claim 2, characterized in that: The length of the sliding time window is also dynamically adjusted according to the construction stage: Excavation stage: The vibration sensor window is shortened to 3-5 minutes, and the earth pressure cell window is shortened to 15-30 minutes; Support stage: The tilt sensor window is extended to 18~24h, and the static level window is extended to 24~48h.

10. The dangerous and major project safety management system according to claim 7, characterized in that: Also includes: A mobile terminal receives the primary and secondary warning signals generated by the zone controller in real time. The mobile terminal marks the boundary of the secondary warning zone on the electronic plan. The boundary of the warning zone extends outward by 3 to 5 meters based on the position of the support structure unit to form a closed polygon. The mobile terminal has built-in: A security control module deploys an electronic fence in the closed polygon, the electronic fence is linked to the access controller, and when the mobile terminal receives both the first-level warning signal and the second-level warning signal, the entrance and exit gates of the warning area are closed; An emergency response module, connected to the on-site sound and light alarm, activates different alarm modes according to the warning level. A level one warning triggers a yellow rotating warning light and an intermittent buzzer, while a level two warning triggers a red strobe warning light and a continuous buzzer; Among them, the sound and light alarm is synchronously encoded with the micro linear motor embedded in the lining of the safety helmet. When the first-level warning is triggered, the micro linear motor vibrates intermittently at a frequency of 2Hz, and when the second-level warning is triggered, the micro linear motor vibrates continuously at a frequency of 5Hz.

Images