Tunnel Gas Risk Prediction and Adaptive Alarm System Based on Big Data from the Construction Industry

Abstract

This invention relates to the field of risk warning technology, specifically disclosing an adaptive alarm system for tunnel gas risk prediction based on big data from the construction industry. The system includes: a cavity identification module that reads curvature and slope information from tunnel BIM data to construct an airflow stagnation cavity; a gradient analysis module that uses the stagnation cavity as boundary conditions to extract the pressure sealing strength on both sides of the shear layer, calculates the mass transfer resistance coefficient, and establishes a gas concentration accumulation gradient field by combining the voxel unit burial distance; an evolution analysis module that maps the real-time location of workers to the gradient field, quantifies the cumulative exposure dose, and generates dynamic evolution values; an evacuation analysis module that triggers an early warning based on the dynamic evolution values ​​and extracts the shortest evacuation vector; and a contamination analysis module that determines the turbulent contamination mixing effect, performs counter-current bias compensation on the evacuation vector, and generates adaptive evacuation guidance commands. This invention is beneficial for achieving early warning of gas risks and reducing personnel exposure risks.

Description

Technical Field

[0001] This invention relates to the field of risk warning technology, specifically to an adaptive alarm system for tunnel gas risk prediction based on big data from the construction industry. Background Technology

[0002] Tunnel construction is a crucial part of transportation infrastructure. During tunnel construction, geometric distortions such as changes in the curvature of the design axis and changes in the longitudinal slope are common. These areas are prone to forming airflow stagnation cavities where airflow is obstructed. The gas inside these cavities cannot exchange with the main airflow, which can easily lead to the accumulation of harmful or flammable gases, becoming a potential hazard to the safety of construction workers.

[0003] Traditional tunnel gas risk monitoring and early warning systems fail to dynamically quantify exposure risks by combining real-time worker work trajectories and dwell times, leading to delayed or misjudgments in early warnings. Furthermore, evacuation guidance after an early warning is based on fixed path planning, failing to consider the turbulent entrainment and mixing effects that may exist in the fluid shear layer. The traditional shortest path may cross secondary high-risk mixing zones, exposing workers to additional safety risks during evacuation. These shortcomings make it difficult to predict gas risks in tunnel construction in advance, resulting in a mismatch between early warning signals and actual worker exposure conditions. Evacuation guidance also lacks adaptability to complex flow field environments, reducing the effectiveness of gas safety protection.

[0004] Therefore, this invention provides an adaptive alarm system for tunnel gas risk prediction based on big data from the construction industry. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive alarm system for tunnel gas risk prediction based on big data from the construction industry, in order to solve the aforementioned background problems.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A tunnel gas risk prediction and adaptive alarm system based on big data from the construction industry includes the following modules:

[0008] Gradient analysis module: used to obtain the airflow stagnation cavity and use it as boundary conditions, and extract the pressure sealing strength of the fluid shear layer by combining the tangential velocity of the main airflow; based on the pressure sealing strength, gas stagnation analysis is performed to obtain the mass transfer stagnation coefficient of gas passing through the shear layer, and a cumulative gradient field of gas concentration is established.

[0009] Evolution Analysis Module: This module maps the worker's real-time location coordinates to the cumulative gradient field of gas concentration, quantifies the worker's cumulative exposure dose within the cumulative gradient field, and generates dynamic evolution values ​​of the worker's cumulative exposure dose.

[0010] Evacuation Analysis Module: Used to determine the early warning trigger based on dynamic evolution values. If an early warning is triggered, the direction vector from the worker's current coordinates to the fluid shear layer separation boundary is extracted based on the airflow stagnation cavity and used as the shortest evacuation vector.

[0011] Entrainment Analysis Module: Based on the shortest evacuation vector, the Reynolds number of the main airflow is obtained and it is determined whether there is a turbulent entrainment mixing effect at the fluid shear layer interface. If so, the shortest evacuation vector is compensated for by reverse flow bias, and an adaptive evacuation guidance command is generated.

[0012] Furthermore, the method for obtaining the airflow stagnation cavity is as follows:

[0013] Obtain curvature and slope data from the tunnel BIM and identify geometric distortion to obtain the geometric distortion zone; perform spatial calculation analysis on the geometric distortion zone to establish the airflow stagnation cavity of the dead air.

[0014] Furthermore, the spatial solution analysis is performed as follows:

[0015] Obtain the Reynolds number and radius of curvature of the main airflow, call the preset flow field correction table of the construction industry big data, and search for the corresponding separation angle in the flow field correction table based on the radius of curvature and Reynolds number as index parameters;

[0016] Based on the separation angle obtained by matching, combined with the tunnel center coordinates corresponding to the geometric distortion zone of BIM data, the three-dimensional spatial coordinates of the fluid shear layer separation point are calculated.

[0017] The characteristic height of the geometric distortion region is extracted to calculate the reattachment length, thus obtaining the reattachment length of the airflow;

[0018] Starting from the three-dimensional spatial coordinates of the fluid shear layer separation point, the reattachment length is extended along the main airflow direction of the tunnel design axis to determine the three-dimensional spatial coordinates of the airflow reattachment point.

[0019] Construct a streamlined boundary surface connecting the fluid shear layer separation point and the airflow reattachment point;

[0020] The closed area enclosed between the streamlined boundary surface and the tunnel inner wall corresponding to the geometric distortion zone is used as the airflow stagnation cavity of the airflow dead zone.

[0021] Furthermore, the method for performing the gas retardation analysis is as follows:

[0022] A standard reference dynamic pressure threshold is set for tunnel ventilation design. The calculated pressure sealing strength is then compared with the standard reference dynamic pressure threshold to obtain the relative pressure strength, which is used as the mass transfer retardation coefficient.

[0023] Furthermore, the method for obtaining the pressure sealing strength is as follows:

[0024] The airflow stagnation cavity is meshed and voxelized to generate a voxel set containing boundary coordinate data;

[0025] Based on voxel sets, the external surface of the airflow stagnation cavity in contact with the main airflow is identified and extracted as the shear layer boundary interface.

[0026] The average wind speed is obtained and then projected onto the tangential direction of the shear layer interface using the vector projection method to obtain the tangential velocity of the main airflow.

[0027] Calculate the product of the square of the tangential velocity and the air density, and define half of the product as the dynamic pressure of the fluid shear layer;

[0028] The dynamic pressure value is used as the pressure sealing strength of the fluid shear layer from the outside to the inside.

[0029] Furthermore, the dynamic evolution value is obtained in the following way:

[0030] Extract the simulated concentration value of the target voxel unit as the instantaneous exposure concentration of the worker at the current moment;

[0031] Set a time sampling step and continuously collect instantaneous exposure concentrations along the worker's trajectory according to the time sampling step;

[0032] Calculate the product of the instantaneous exposure concentration at the current moment and the time sampling step size to obtain the exposure increment within a single step at the current moment;

[0033] The current exposure increment is summed with the exposure increment recorded in the previous time step, and the cumulative total value of the current exposure increment is updated in real time.

[0034] The cumulative total of the exposure increments calculated in real time is used as the dynamic evolution value of the cumulative exposure dose.

[0035] Furthermore, the method for extracting the shortest evacuation vector is as follows:

[0036] Traverse the coordinates of all boundary voxels that constitute the fluid shear layer separation boundary;

[0037] Calculate the Euclidean distance between the worker's current coordinates and the coordinates of each boundary voxel inside the airflow stagnation cavity, and select the boundary voxel coordinates with the smallest Euclidean distance as the ideal evacuation landing point.

[0038] Construct a direction vector from the worker's current coordinates to the ideal evacuation landing point, and define the direction vector as the shortest evacuation vector.

[0039] Furthermore, the method for determining whether turbulent entrainment mixing effects exist is as follows:

[0040] Obtain the solid mixing thickness and construct a secondary high-risk mixing zone containing the solid volume;

[0041] Perform geometric collision detection on the shortest evacuation vector to determine whether the path segment of the shortest evacuation vector passes through the secondary high-risk mixing zone and whether the gas concentration exceeds the limit;

[0042] If the gas does not pass through or the gas concentration does not exceed the limit, the shortest evacuation vector is directly marked as the final adaptive evacuation guidance command.

[0043] If the worker passes through a secondary high-risk mixing zone and the gas concentration exceeds the limit, it is determined that there is a turbulent entrainment mixing effect. The relative area of ​​the worker's current position in the airflow retention cavity is identified, and countercurrent bias compensation is performed based on the relative area.

[0044] Furthermore, the method for obtaining the physical mixing thickness is as follows:

[0045] Obtain the three-dimensional spatial coordinates of the fluid shear layer separation point, and the average wind speed corresponding to the three-dimensional spatial coordinates;

[0046] Calculate the projection point of the worker's current coordinates on the fluid shear layer separation boundary, and obtain the flow distance along the tunnel axis between the projection point and the fluid shear layer separation point;

[0047] Based on the mixing length theory of fluid mechanics, and combined with the average wind speed and flow direction distance, the solid mixing thickness of the fluid shear layer in the downstream direction is calculated.

[0048] Furthermore, the method for performing the reverse bias compensation is as follows:

[0049] Obtain the worker's current real-time location coordinates;

[0050] When the reverse flow offset compensation is triggered, a fixed safety deflection angle is set. Using the worker's current real-time positioning coordinates as the rotation origin, the shortest evacuation vector is rotated in the direction of the main airflow source by the safety deflection angle to generate the final adaptive evacuation guidance command.

[0051] The beneficial effects of this invention are as follows:

[0052] 1. By extracting cross-sections from tunnel BIM data and considering the radius of curvature and longitudinal slope, geometric distortion zones are identified using comparison rules. Simultaneously, Reynolds number is calculated by integrating main ventilation fan wind speed data and aerodynamic parameters. A flow field correction table based on historical simulation data is used to match the separation angle. Finally, airflow stagnation cavities are constructed through feature height extraction and reattachment length formula calculation. This approach, combining BIM technology, fluid mechanics principles, and big data from the construction industry, helps to accurately reflect the actual geometry of the tunnel and airflow patterns, thus identifying stagnant water zones where air accumulation occurs.

[0053] 2. Using the airflow stagnation cavity as the boundary, the shear layer boundary interface is extracted through mesh voxelization to obtain the pressure closure strength corresponding to the tangential velocity of the main airflow. The mass transfer resistance coefficient is then obtained by comparing it with the standard reference dynamic pressure threshold, which helps to intuitively reflect the relative strength of the shear layer's blockade of the gas. Subsequently, the simulated concentration is calculated by combining the burial distance of the voxel elements and the mass transfer resistance coefficient. A continuous concentration accumulation gradient field is generated through spatial interpolation, including both the concentration distribution and the set of direction vectors. This helps to present the spatial distribution differences and accumulation trends of the gas within the stagnation cavity.

[0054] 3. The worker's real-time positioning coordinates are transformed to a coordinate system consistent with the BIM model, mapped to the concentration accumulation gradient field, and the target voxel unit is locked. Different exposure concentration benchmarks are set inside and outside the cavity to reduce data interference from irrelevant areas. Instantaneous exposure concentrations are collected by sampling at fixed time steps, and dynamic evolution values ​​are generated by incremental superposition. This can naturally distinguish the exposure differences between different work modes such as rapid passage and static stay. Through dynamic quantification, it is beneficial to track the worker's cumulative exposure at different spatial locations in real time, forming a dose monitoring data stream that changes over time, and achieving matching of risk assessment with the worker's actual work trajectory and stay status.

[0055] 4. Early warning is triggered by comparing the dynamic evolution values ​​of cumulative exposure dose, ensuring that the warning activation matches the actual exposure risk status of workers. By traversing the voxel coordinates of the fluid shear layer separation boundary, the shortest Euclidean distance between the worker's current position and the boundary voxel is calculated, ideal evacuation landing points are selected, and the shortest evacuation vector is generated. This provides workers with an evacuation direction reference that fits the site's spatial layout, allowing evacuation path planning to fully incorporate the boundary characteristics of the airflow stagnation cavity. A secondary high-risk mixing zone is constructed, and the existence of turbulent entrainment mixing effects is determined through geometric collision detection. When turbulent entrainment effects exist, the evacuation vector is compensated for with a counter-current offset to reduce the additional risks brought by the mixing zone. Simultaneously, evacuation is guided by azimuth vibration feedback from intelligent terminals, and closed-loop monitoring is conducted to ensure workers do not cross the shear layer interface. This allows evacuation guidance to dynamically adapt to changes in the flow field, improving the safety and efficiency of the worker evacuation process. Attached Figure Description

[0056] The invention will now be further described with reference to the accompanying drawings.

[0057] Figure 1 This is a functional module diagram of the tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry, which is a functional module diagram of the present invention.

[0058] Figure 2 This is a flowchart illustrating the determination process for triggering an early warning signal in this invention;

[0059] Figure 3This is a flowchart of the adaptive alarm method for tunnel gas risk prediction based on big data in the construction industry in this invention. Detailed Implementation

[0060] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0061] Example 1: As Figure 1 As shown, the tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry includes the following modules:

[0062] Cavity recognition module: used to acquire curvature and slope point data in tunnel BIM and perform geometric distortion recognition to obtain geometric distortion areas; perform spatial calculation and analysis on geometric distortion areas to establish airflow stagnation cavities with stagnant water.

[0063] The method for obtaining curvature and slope point data from tunnel BIM and performing geometric distortion identification to determine the geometric distortion zone is as follows:

[0064] Preferably, the BIM (Building Information Modeling) data of the current construction section of the tunnel is read through a standard data interface, and the tunnel design axis is extracted from the BIM data;

[0065] Based on the tunnel design axis, a preset step length is defined and the tunnel cross section is cut according to the preset step length. The radius of curvature and longitudinal slope of each cross section are extracted.

[0066] Preferably, the preset step size is 1 meter;

[0067] Calculate the rate of change of the radius of curvature of adjacent cross sections, and set the comparison rules for geometric distortion;

[0068] If the rate of change of the radius of curvature of adjacent cross sections and the longitudinal slope satisfy the comparison rules of geometric distortion, then adjacent cross sections are grouped into a segment as a geometric distortion zone.

[0069] For example, the comparison rules for geometric distortion can be set as follows:

[0070] According to Rule 1, the rate of change of the radius of curvature is greater than 0.05;

[0071] Rule 2: The longitudinal slope of adjacent cross sections changes from positive to negative;

[0072] If either comparison rule one or comparison rule two is satisfied, then the rate of change of the radius of curvature of the adjacent cross section and the longitudinal slope are determined to satisfy the comparison rule of geometric distortion.

[0073] Among them, the method of establishing the airflow stagnation cavity of the airflow dead water by performing spatial solution analysis on the geometric distortion region is as follows:

[0074] The system interface is used to read the wind speed value of the tunnel main ventilation fan's wind speed sensor at the sampling point and define a time window;

[0075] The time window includes N1 sampling points;

[0076] The average wind speed is calculated as the mean wind speed within the time window.

[0077] Simultaneously, the hydraulic diameter of the corresponding cross section is read from the BIM data to obtain the standard aerodynamic viscosity and air density;

[0078] Calculate the product of air density, average wind speed, and hydraulic diameter, and then ratio the product to air viscosity to obtain the Reynolds number of the main airflow.

[0079] The system calls the preset flow field correction table of the construction industry big data, and searches for the corresponding separation angle in the flow field correction table based on the radius of curvature and Reynolds number as index parameters.

[0080] It should be noted that the flow field correction table is a lookup table built based on historical CFD (Computational Fluid Dynamics) simulation data. The lookup table consists of multiple data sets: [Reynolds number interval, radius of curvature, separation angle].

[0081] For example, the flow field correction table is shown in Table 1:

[0082] Table 1: Flow Field Correction Table

[0083]

[0084] The separation angle is the angle between the tunnel cross-section as the center and the direction of the main airflow.

[0085] Based on the separation angle obtained by matching, and combined with the tunnel center coordinates corresponding to the geometric distortion zone of BIM data, the three-dimensional spatial coordinates of the fluid shear layer separation point are calculated.

[0086] As will be understood by those skilled in the art, the coordinate calculation is performed as follows:

[0087] The coordinates of the center point of the current cross-section of the tunnel are obtained as the origin of the ray; the separation angle is deflected towards the geometric distortion zone (i.e., the inside of the bend or the slope change side) with the main airflow direction as the reference axis to construct the separation direction ray in three-dimensional space; the ray intersection function of the BIM software standard is called to calculate the first geometric intersection point between the separation direction ray and the inner wall mesh of the tunnel BIM model, and the three-dimensional coordinates of the geometric intersection point are extracted to realize coordinate solution processing;

[0088] Extract the characteristic height of the geometric distortion zone, call the preset reattachment length calculation formula, and calculate the reattachment length of the airflow based on the Reynolds number and characteristic height of the main airflow.

[0089] For example, the method for extracting the feature height of the geometric distortion zone is as follows: traverse all adjacent cross sections contained within the geometric distortion zone; if the geometric distortion zone is triggered by comparison rule one, compare the rate of change of the radius of curvature of each adjacent cross section within the segment, select the adjacent cross section with the largest rate of change, and define the intersection of adjacent cross sections as the curvature change location; based on BIM data, calculate the geometric deviation distance between the preceding and subsequent cross sections at the curvature change location in the tangential direction of the tunnel inner wall, and use it as the feature height triggered by comparison rule one;

[0090] If the geometric distortion zone is triggered by comparison rule two, then the adjacent cross sections whose longitudinal slope changes from positive to negative are selected, and the boundary between the adjacent cross sections is defined as the slope reversal position; based on BIM data, the absolute value of the vertical elevation difference between the center point of the preceding cross section and the center point of the following cross section at the slope reversal position is calculated, and the absolute value is used as the maximum vertical elevation difference, which is used as the characteristic height triggered by comparison rule two;

[0091] It should be noted that if the geometric distortion area satisfies both comparison condition one and comparison condition two, the feature height extraction method of comparison rule two will be given priority because the influence of the longitudinal slope change on airflow retention is higher.

[0092] For example, the formula for calculating the reattachment length is: ,in, For reattachment length, For feature height, The Reynolds number of the dominant flow, It is the natural logarithm. and This is a preset proportionality constant;

[0093] in, The value range is 4.0 to 5.0 (preferably 4.5). The value range is from 0.4 to 0.6 (preferably 0.5). and The regression analysis of 30 sets of tunnel CFD simulation data using the least squares method yielded the optimal values ​​C1=4.5 and C2=0.5, which are suitable for circular tunnel cross sections.

[0094] Starting from the three-dimensional spatial coordinates of the fluid shear layer separation point, the reattachment length is extended along the main airflow direction of the tunnel design axis to determine the three-dimensional spatial coordinates of the airflow reattachment point.

[0095] Construct a streamlined boundary surface connecting the fluid shear layer separation point and the airflow reattachment point;

[0096] The closed area enclosed between the streamlined boundary surface and the tunnel inner wall corresponding to the geometric distortion zone is used as the airflow stagnation cavity of the airflow dead zone.

[0097] The gradient analysis module is used to extract the pressure sealing strength of the fluid shear layer by taking the airflow stagnation cavity as the boundary condition and combining it with the tangential velocity of the main airflow; based on the pressure sealing strength, gas stagnation analysis is performed to obtain the mass transfer stagnation coefficient of gas passing through the shear layer, and a cumulative gradient field of gas concentration is established.

[0098] The method of extracting the pressure sealing strength of the fluid shear layer by using the airflow stagnation cavity as a boundary condition and combining it with the tangential velocity of the main airflow is as follows:

[0099] The airflow stagnation cavity is meshed and voxelized to generate a voxel set containing boundary coordinate data;

[0100] Based on voxel sets, the external surface of the airflow stagnation cavity in contact with the main airflow is identified and extracted as the shear layer boundary interface.

[0101] The average wind speed is obtained and then projected onto the tangential direction of the shear layer interface using the vector projection method to obtain the tangential velocity of the main airflow.

[0102] Based on Bernoulli's principle in fluid mechanics, the product of the square of the tangential velocity and the air density is calculated, and half of the product is defined as the dynamic pressure of the fluid shear layer.

[0103] The dynamic pressure value is used as the pressure sealing strength of the fluid shear layer from the outside to the inside.

[0104] It should be noted that the aerodynamic cover effect formed by high-speed fluid means that the faster the tangential velocity of the main airflow, the greater the dynamic pressure formed at the interface, the stronger the sealing effect on the gas inside the dead water zone, and the greater the corresponding pressure sealing strength value.

[0105] The method for obtaining the mass transfer resistance coefficient of gas passing through the shear layer based on gas retardation analysis using pressure sealing strength is as follows:

[0106] A standard reference dynamic pressure threshold is set for tunnel ventilation design. The calculated pressure sealing strength is compared with the standard reference dynamic pressure threshold to obtain the dimensionless relative pressure strength, which is used as the mass transfer retardation coefficient.

[0107] It should be noted that the dynamic pressure value of the main airflow at the design maximum wind speed is obtained as the preset standard reference dynamic pressure threshold; the mass transfer retardation coefficient is used to characterize the degree of inhibition of the outward diffusion of gas in the cavity by the shear layer. The larger the value, the lower the proportion of gas in the cavity diffusing towards the main airflow side per unit time. Therefore, under steady-state or quasi-steady-state conditions, it can be equivalent to the amplification factor of the peak concentration in the cavity.

[0108] The method for establishing the cumulative gradient field of gas concentration is as follows:

[0109] Traverse all voxel units inside the airflow stagnation cavity of the dead airflow, calculate the shortest Euclidean distance from the center coordinates of each voxel unit to the shear layer interface, and use it as the deep burial distance of the voxel unit.

[0110] Concentration enrichment analysis was performed based on the burial distance and mass transfer retardation coefficient to obtain the simulated concentration value of each voxel unit.

[0111] The method for concentration enrichment analysis is as follows:

[0112] The concentration at the shear layer interface where the airflow stagnation cavity contacts the main airflow is set as the tunnel background concentration. ;

[0113] The peak gas concentration from historical engineering data in the construction industry's big data database is used, multiplied by the calculated mass transfer resistance coefficient, to determine the estimated maximum accumulation concentration at the deepest part of the cavity. ;

[0114] Traverse the voxel units within the cavity, calculate the distance d from the current voxel unit to the shear layer interface, and obtain the maximum distance from the cavity interior to the shear layer interface. ;

[0115] Using formula Calculate the simulated concentration value of the current voxel unit. ;

[0116] A three-dimensional data matrix containing all voxel units and their corresponding simulated concentration values ​​is constructed. A spatial interpolation algorithm is used to smoothly connect the discrete voxel concentration values ​​in the matrix to generate a continuous concentration distribution cloud map.

[0117] Calculate the concentration growth rate along the spatial coordinate axis at each point in the distribution cloud map, and generate a set of direction vectors pointing to the high-concentration core area;

[0118] The spatial model containing a continuous concentration distribution and a set of direction vectors is defined as the cumulative gradient field of gas concentration;

[0119] It should be noted that gases or hazardous gases refer to various toxic, harmful, flammable and explosive gases that may be generated or accumulated during tunnel construction, including: methane that overflows from the strata, nitrogen oxides (NOx) and carbon monoxide (CO) produced by blasting, etc.

[0120] The purpose of establishing the cumulative gradient field of gas concentration is as follows:

[0121] Objective 1: To transform discrete voxel simulation data into a continuous spatial distribution model, so that any coordinate point within the tunnel has a queryable concentration estimate.

[0122] Objective 2: Establishing a gradient field helps to obtain the numerical value of the concentration, and at the same time, the location of the core accumulation area can be identified by the direction of the gradient, that is, the direction of the fastest concentration increase.

[0123] Example 2: Please refer to Figure 1 As shown, the tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry includes the following modules:

[0124] Evolution Analysis Module: This module maps the worker's real-time location coordinates to the cumulative gradient field of gas concentration, quantifies the worker's cumulative exposure dose within the cumulative gradient field, and generates dynamic evolution values ​​of the worker's cumulative exposure dose.

[0125] The method for mapping the worker's real-time location coordinates to the cumulative gradient field of gas concentration is as follows:

[0126] Preferably, the real-time three-dimensional spatial coordinates of workers are obtained through a personnel positioning system, and then the real-time three-dimensional spatial coordinates are converted to a coordinate system consistent with the tunnel BIM model;

[0127] Establish a spatial index for the cumulative gradient field of gas concentration, and lock the voxel unit containing real-time three-dimensional spatial coordinates as the target voxel unit.

[0128] The method for quantifying the cumulative exposure dose of workers located within the cumulative gradient field and generating dynamic evolution values ​​of workers' cumulative exposure dose is as follows:

[0129] Extract the simulated concentration value of the target voxel unit as the instantaneous exposure concentration of the worker at the current moment;

[0130] It should be noted that the system will first determine whether the worker's coordinates are inside the airflow stagnation chamber in the dead zone. If the worker is outside, the instantaneous exposure concentration is set to the safe background value by default. Only when the worker enters the airflow stagnation chamber will the system start reading the high concentration values ​​in the gradient field.

[0131] Set a time sampling step and continuously collect instantaneous exposure concentrations along the worker's trajectory according to the time sampling step;

[0132] Calculate the product of the instantaneous exposure concentration at the current moment and the time sampling step size to obtain the exposure increment within a single step at the current moment;

[0133] The current exposure increment is summed with the exposure increment recorded in the previous time step, and the cumulative total value of the current exposure increment is updated in real time.

[0134] It should be noted that the cumulative overlay calculation can automatically distinguish between two behavior modes: rapid passage and stationary lingering. If a worker quickly passes through a high-concentration area, the cumulative total value is low due to fewer overlays. However, if a worker performs peripheral work or rests in a high-concentration area, the number of overlays increases continuously as the dwell time lengthens, and the cumulative total value rises rapidly. The system continuously monitors the worker's location. If it detects that a worker has entered a safe area (i.e., the environmental concentration is below the background threshold) and remains there for more than a preset metabolic recovery period (e.g., 30 minutes), the cumulative exposure increment is exponentially decayed. This is used to simulate the human body's natural metabolic process of harmful gases, reducing false high-risk alarms caused by long-term accumulation.

[0135] The cumulative total value of the exposure increment obtained by real-time superposition calculation is defined as the dynamic evolution value of the cumulative exposure dose, forming a dose monitoring data stream that monotonically increases over time.

[0136] Evacuation Analysis Module: Used to determine the early warning trigger based on dynamic evolution values. If an early warning is triggered, the direction vector from the worker's current coordinates to the fluid shear layer separation boundary is extracted based on the airflow stagnation cavity and used as the shortest evacuation vector.

[0137] The method for determining the early warning trigger for dynamic evolution values ​​is as follows:

[0138] like Figure 2 As shown, a safety threshold corresponding to the dynamic evolution value is set. If the dynamic evolution value of the worker's cumulative exposure dose is higher than or equal to the set safety threshold, an early warning signal is triggered by the system.

[0139] It should be noted that the safety threshold is set based on the tunnel construction safety technical specification standard: it is based on the 8-hour time-weighted average permissible exposure limit of the target harmful gas (such as CO, methane), and is converted according to "cumulative exposure dose = concentration × time" (e.g., permissible limit for CO). The corresponding security threshold = In confined space operations within tunnels, the threshold is lowered by 20% to reserve safety redundancy and match the actual operational risks.

[0140] If the dynamic evolution value is lower than the set safety threshold, the change in the dynamic evolution value will be continuously monitored.

[0141] The method for extracting the direction vector from the worker's current coordinates to the fluid shear layer separation boundary, based on the airflow stagnation cavity, is as follows:

[0142] Traverse the coordinates of all boundary voxels that constitute the fluid shear layer separation boundary;

[0143] Calculate the Euclidean distance between the worker's current coordinates and the coordinates of each boundary voxel inside the airflow stagnation cavity, and select the boundary voxel coordinates with the smallest Euclidean distance as the ideal evacuation landing point.

[0144] Construct a direction vector from the worker's current coordinates to the ideal evacuation landing point, and define the direction vector as the shortest evacuation vector;

[0145] It should be noted that the shortest evacuation vector represents the shortest physical path required for a worker to escape from the airflow congestion cavity under ideal conditions, without considering secondary disasters in the flow field or personnel congestion.

[0146] Entrainment Analysis Module: Based on the shortest evacuation vector, obtain the Reynolds number of the main airflow and determine whether there is a turbulent entrainment mixing effect at the fluid shear layer interface. If so, perform countercurrent bias compensation on the shortest evacuation vector and generate adaptive evacuation guidance commands.

[0147] The method for obtaining the Reynolds number of the main airflow and determining whether turbulent entrainment mixing effects exist at the fluid shear layer interface based on the shortest withdrawal vector is as follows:

[0148] Obtain the three-dimensional spatial coordinates of the fluid shear layer separation point, and the average wind speed corresponding to the three-dimensional spatial coordinates;

[0149] Calculate the projection point of the worker's current coordinates on the fluid shear layer separation boundary, and obtain the flow distance along the tunnel axis between the projection point and the fluid shear layer separation point;

[0150] Based on the mixing length theory of fluid mechanics, and combined with the average wind speed and flow direction distance, the solid mixing thickness of the fluid shear layer in the downstream direction is calculated.

[0151] For example, through the formula: Computational fluid shear layer thickness in the downstream direction of solid mixing ;

[0152] in, To obtain the average wind speed, The average flow velocity within the stagnant airflow cavity of the dead airflow system. For flow direction distance, The preset empirical coefficient for turbulence growth (within the range of 0.1-0.2, preferably 0.15);

[0153] Using the fluid shear layer separation boundary as the central reference plane and the solid mixing thickness as the expansion width on both sides, a secondary high-risk mixing zone containing the solid volume is constructed.

[0154] Perform geometric collision detection on the shortest evacuation vector to determine whether the path segment of the shortest evacuation vector passes through the secondary high-risk mixing zone and whether the gas concentration exceeds the limit;

[0155] Preferably, the geometric collision detection method is as follows: if the path segment of the shortest evacuation vector passes through the secondary high-risk mixing zone, and the path length in the mixing zone is greater than 0 and the gas concentration exceeds the limit, then it is determined from the physical space that there is a turbulent entrainment mixing effect at the fluid shear layer separation boundary.

[0156] It should be noted that exceeding the gas concentration limit refers to the simulated concentration value in the associated concentration cumulative gradient field. A dangerous concentration threshold is set within the mixing zone. If the gas concentration is higher than or equal to the dangerous concentration threshold, it is determined that the gas concentration exceeds the limit; otherwise, the gas concentration is continuously monitored.

[0157] If the path segment of the shortest evacuation vector does not pass through the secondary high-risk mixing zone or the gas concentration does not exceed the limit, it is determined that there is no turbulent entrainment mixing effect.

[0158] If it is determined that there is no turbulent entrainment mixing effect, the shortest evacuation vector is directly marked as the final adaptive evacuation guidance command.

[0159] If the worker passes through a secondary high-risk mixing zone and the gas concentration exceeds the limit, it is determined that there is a turbulent entrainment mixing effect, and the relative area of ​​the worker's current position in the airflow stagnation cavity is identified.

[0160] The geometric center point of the airflow stagnation cavity along the tunnel axis is used as the boundary. The upstream side is the area near the fluid shear layer separation point, and the downstream side is the area near the airflow reattachment point.

[0161] If the worker is located in the upstream area of ​​the main airflow in the airflow stagnation cavity, the reverse flow offset compensation is triggered because the path loss of the offset evacuation is low.

[0162] If the worker is located downstream of the main airflow in the airflow stagnation cavity, the system first calculates the peak gas concentration on the shortest evacuation vector path. If the peak exceeds the preset instantaneous concentration threshold, the system will forcibly trigger the reverse flow offset compensation to guide the worker to detour upstream. Otherwise, the reverse flow offset compensation will not be triggered, and the shortest evacuation vector will be marked as the final adaptive evacuation guidance command.

[0163] The method for generating the final adaptive evacuation guidance command by performing reverse current offset compensation on the shortest evacuation vector is as follows:

[0164] For example, the execution and masking of the reverse bias compensation logic are based on the following:

[0165] It should be noted that when the worker is on the upstream side (close to the separation point), although the fluid shear layer is thinner and the worker is close to the clean main airflow area upstream, implementing counter-current offset and fine-tuning the evacuation angle at this time will result in a small increase in the actual evacuation path length of the worker, which is beneficial to reducing the risk of turbulent entrainment in the shear layer.

[0166] When workers are downstream (near the reattachment point), the shear layer is fully developed and relatively thick. The system first determines the safety of the path: if the gas concentration on the calculated path does not exceed the limit, the shortest exposure time principle is followed, the bias logic is shielded, and workers are guided to evacuate vertically; if the path concentration exceeds the instantaneous concentration threshold, the shortest path principle is broken, the reverse bias is forcibly triggered, and workers are guided to detour upstream to avoid the high-risk area;

[0167] The system has a pre-set table of instantaneous concentration threshold values ​​for common tunnel risk gases (such as carbon monoxide, hydrogen sulfide, and methane) (e.g., 1200 ppm for carbon monoxide and 100 ppm for hydrogen sulfide). During system initialization, the system automatically indexes and loads the corresponding values ​​from the table based on the main risk gas types identified in the geological survey report of the current tunnel construction section, using these values ​​as the criteria for safety assessment.

[0168] When the reverse flow offset compensation is triggered, a fixed safety deflection angle (e.g., 30 to 45 degrees) is set. The worker's current real-time positioning coordinates are used as the rotation origin. The shortest evacuation vector is rotated in the direction of the main airflow source by the safety deflection angle to generate the final adaptive evacuation guidance command.

[0169] Mark the rotation-corrected vector as the final adaptive withdrawal guidance command;

[0170] The purpose of determining the turbulent entrainment mixing effect and performing adaptive path compensation is as follows:

[0171] Objective 1: To correct the shortcomings of traditional algorithms that only consider the shortest geometric distance and ignore the risks of fluid dynamics, and to prevent workers from taking shortcuts but directly crossing secondary high-risk mixing zones;

[0172] Objective 2: To weigh the trade-off between evacuation distance and the risk of gas inhalation, depending on the worker's specific location within the cavity (upstream or downstream).

[0173] The methods for guiding workers to implement a closed-loop evacuation are as follows:

[0174] The final adaptive evacuation guidance command is parsed into azimuth vibration signals and sent to the smart terminal worn by the worker. Vibration feedback guides the worker's movement direction.

[0175] Continuously monitor the real-time positioning coordinates of workers and calculate the positional relationship between workers and the fluid shear layer interface in real time;

[0176] If it is determined that the worker's real-time positioning coordinates have crossed the fluid shear layer interface and are located outside the airflow stagnation cavity, then the sending of guidance commands will stop and the alarm closed loop for the worker will be released.

[0177] Example 3: Please refer to Figure 3As shown, the adaptive alarm method for tunnel gas risk prediction based on big data in the construction industry includes the following steps:

[0178] S1. Obtain curvature and slope point data from the tunnel BIM and perform geometric distortion identification to obtain the geometric distortion area; perform spatial solution analysis on the geometric distortion area to establish the airflow stagnation cavity of the airflow dead water;

[0179] S2. Using the airflow stagnation cavity as a boundary condition, the pressure sealing strength of the fluid shear layer is extracted in combination with the tangential velocity of the main airflow; gas stagnation analysis is performed based on the pressure sealing strength to obtain the mass transfer stagnation coefficient of gas passing through the shear layer, and a cumulative gradient field of gas concentration is established.

[0180] S3. Map the worker's real-time location coordinates to the cumulative gradient field of gas concentration, quantify the worker's cumulative exposure dose within the cumulative gradient field, and generate the dynamic evolution value of the worker's cumulative exposure dose.

[0181] S4. Determine the early warning trigger for the dynamic evolution value. If the early warning is triggered, extract the direction vector from the worker's current coordinate point to the fluid shear layer separation boundary based on the airflow stagnation cavity, and use it as the shortest evacuation vector.

[0182] S5. Based on the shortest evacuation vector, obtain the Reynolds number of the main airflow and determine whether there is a turbulent entrainment mixing effect at the fluid shear layer boundary. If so, perform countercurrent offset compensation on the shortest evacuation vector to generate an adaptive evacuation guidance command.

[0183] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry, characterized in that: Includes the following modules: Gradient analysis module: used to obtain the airflow stagnation cavity and use it as boundary conditions, and extract the pressure sealing strength of the fluid shear layer by combining the tangential velocity of the main airflow; based on the pressure sealing strength, gas stagnation analysis is performed to obtain the mass transfer stagnation coefficient of gas passing through the shear layer, and a cumulative gradient field of gas concentration is established. The method for performing the gas retardation analysis is as follows: A standard reference dynamic pressure threshold is set for tunnel ventilation design. The calculated pressure sealing strength is compared with the standard reference dynamic pressure threshold to obtain the relative pressure strength, which is used as the mass transfer retardation coefficient. The method for obtaining the pressure sealing strength is as follows: The airflow stagnation cavity is meshed and voxelized to generate a voxel set containing boundary coordinate data; Based on voxel sets, the external surface of the airflow stagnation cavity in contact with the main airflow is identified and extracted as the shear layer boundary interface. The average wind speed is obtained and then projected onto the tangential direction of the shear layer interface using the vector projection method to obtain the tangential velocity of the main airflow. Calculate the product of the square of the tangential velocity and the air density, and define half of the product as the dynamic pressure of the fluid shear layer; The dynamic pressure value is used as the pressure sealing strength of the fluid shear layer from the outside to the inside. Evolution Analysis Module: This module maps the worker's real-time location coordinates to the cumulative gradient field of gas concentration, quantifies the worker's cumulative exposure dose within the cumulative gradient field, and generates dynamic evolution values ​​of the worker's cumulative exposure dose. Evacuation Analysis Module: Used to determine the early warning trigger based on dynamic evolution values. If an early warning is triggered, the direction vector from the worker's current coordinates to the fluid shear layer separation boundary is extracted based on the airflow stagnation cavity and used as the shortest evacuation vector. Entrainment Analysis Module: Based on the shortest evacuation vector, the Reynolds number of the main airflow is obtained and it is determined whether there is a turbulent entrainment mixing effect at the fluid shear layer interface. If so, the shortest evacuation vector is compensated for by reverse flow bias, and an adaptive evacuation guidance command is generated.

2. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 1, characterized in that: The method for obtaining the airflow stagnation cavity is as follows: Obtain curvature and slope data from the tunnel BIM and identify geometric distortion to obtain the geometric distortion zone; perform spatial calculation analysis on the geometric distortion zone to establish the airflow stagnation cavity of the dead air.

3. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 2, characterized in that: The method for performing the spatial solution analysis is as follows: Obtain the Reynolds number and radius of curvature of the main airflow, call the preset flow field correction table of the construction industry big data, and search for the corresponding separation angle in the flow field correction table based on the radius of curvature and Reynolds number as index parameters; Based on the separation angle obtained by matching, combined with the tunnel center coordinates corresponding to the geometric distortion zone of BIM data, the three-dimensional spatial coordinates of the fluid shear layer separation point are calculated. The characteristic height of the geometric distortion region is extracted to calculate the reattachment length, thus obtaining the reattachment length of the airflow; Starting from the three-dimensional spatial coordinates of the fluid shear layer separation point, the reattachment length is extended along the main airflow direction of the tunnel design axis to determine the three-dimensional spatial coordinates of the airflow reattachment point. Construct a streamlined boundary surface connecting the fluid shear layer separation point and the airflow reattachment point; The closed area enclosed between the streamlined boundary surface and the tunnel inner wall corresponding to the geometric distortion zone is used as the airflow stagnation cavity of the airflow dead zone.

4. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 1, characterized in that: The dynamic evolution value is obtained as follows: Extract the simulated concentration value of the target voxel unit as the instantaneous exposure concentration of the worker at the current moment; Set a time sampling step and continuously collect instantaneous exposure concentrations along the worker's trajectory according to the time sampling step; Calculate the product of the instantaneous exposure concentration at the current moment and the time sampling step size to obtain the exposure increment within a single step at the current moment; The current exposure increment is summed with the exposure increment recorded in the previous time step, and the cumulative total value of the current exposure increment is updated in real time. The cumulative total of the exposure increments calculated in real time is used as the dynamic evolution value of the cumulative exposure dose.

5. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 1, characterized in that: The method for extracting the shortest evacuation vector is as follows: Traverse the coordinates of all boundary voxels that constitute the fluid shear layer separation boundary; Calculate the Euclidean distance between the worker's current coordinates and the coordinates of each boundary voxel inside the airflow stagnation cavity, and select the boundary voxel coordinates with the smallest Euclidean distance as the ideal evacuation landing point. Construct a direction vector from the worker's current coordinates to the ideal evacuation landing point, and define the direction vector as the shortest evacuation vector.

6. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 1, characterized in that: The method for determining whether turbulent entrainment mixing effect exists is as follows: Obtain the solid mixing thickness and construct a secondary high-risk mixing zone containing the solid volume; Perform geometric collision detection on the shortest evacuation vector to determine whether the path segment of the shortest evacuation vector passes through the secondary high-risk mixing zone and whether the gas concentration exceeds the limit; If the gas does not pass through or the gas concentration does not exceed the limit, the shortest evacuation vector is directly marked as the final adaptive evacuation guidance command. If the worker passes through a secondary high-risk mixing zone and the gas concentration exceeds the limit, it is determined that there is a turbulent entrainment mixing effect. The relative area of ​​the worker's current position in the airflow retention cavity is identified, and countercurrent bias compensation is performed based on the relative area.

7. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 6, characterized in that: The method for obtaining the physical mixing thickness is as follows: Obtain the three-dimensional spatial coordinates of the fluid shear layer separation point, and the average wind speed corresponding to the three-dimensional spatial coordinates; Calculate the projection point of the worker's current coordinates on the fluid shear layer separation boundary, and obtain the flow distance along the tunnel axis between the projection point and the fluid shear layer separation point; Based on the mixing length theory of fluid mechanics, and combined with the average wind speed and flow direction distance, the solid mixing thickness of the fluid shear layer in the downstream direction is calculated.

8. The tunnel gas risk prediction and adaptive alarm system based on big data in the construction industry according to claim 6, characterized in that: The method for performing the reverse bias compensation is as follows: Obtain the worker's current real-time location coordinates; When the reverse flow offset compensation is triggered, a fixed safety deflection angle is set. Using the worker's current real-time positioning coordinates as the rotation origin, the shortest evacuation vector is rotated in the direction of the main airflow source by the safety deflection angle to generate the final adaptive evacuation guidance command.

Images