A tunnel ventilation control method and device based on environment prediction and a medium

Abstract

This invention discloses a tunnel ventilation control method, equipment, and medium based on environmental prediction, relating to the field of environmental management technology. The method includes: collecting and preprocessing environmental data; generating a pollutant heat map based on a pre-trained pollutant diffusion model; overlaying personnel distribution data with the pollutant heat map and determining the tangential direction of the personnel avoidance path as the airflow direction for evacuation guidance, generating evacuation wind direction data; extracting the speed combination and deflection angle of the tunnel ventilation fan from the evacuation wind direction data and outputting a fan control command set; monitoring the pollutant concentration values ​​in each tunnel ventilation area in real time after directional airflow guidance using a smoke concentration sensor; and comparing the predicted concentration values ​​in the pollutant heat map with the actual pollutant concentration values ​​to verify the tunnel ventilation control effect. This invention achieves environmental prediction and airflow guidance correction in the tunnel ventilation control process through the synergy of pollutant heat map generation and evacuation wind direction data generation.

Description

Technical Field

[0001] This invention relates to the field of environmental management technology, and in particular to a tunnel ventilation control method, equipment and medium based on environmental prediction. Background Technology

[0002] In recent years, with the continuous development of technologies such as tunnel operation environment perception, pollutant diffusion modeling, traffic flow monitoring, and ventilation execution control, environmental prediction and active ventilation regulation for tunnel scenarios have gradually shifted from a single-point monitoring mode to a multi-source data-driven mode. Tunnel ventilation control usually revolves around pollution source location, wind speed and direction acquisition, traffic flow density perception, personnel location identification, and pollution concentration evolution analysis. Combined with diffusion models, path planning methods, and ventilation control strategies, it predicts and regulates the airflow organization and pollutant migration patterns inside the tunnel, promoting the evolution of tunnel ventilation control towards dynamic and refined directions.

[0003] However, the main problem in tunnel ventilation regulation is the lack of a unified linkage mechanism between environmental prediction results, personnel evacuation direction and fan control parameters. This makes it difficult to form a continuous connection between pollutant diffusion trend identification, evacuation guidance airflow organization and fan execution control, which in turn affects the pertinence of pollutant diffusion guidance in the tunnel and the effectiveness of ventilation control effect verification. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a tunnel ventilation control method based on environmental prediction to solve the problems of deficiencies in prediction accuracy, control coordination and dynamic optimization.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a tunnel ventilation control method based on environmental prediction, comprising: collecting environmental data and preprocessing it; the environmental data including pollution source coordinates, traffic density data, wind speed and direction data within the tunnel, and personnel distribution data; based on a pre-trained pollutant diffusion model, predicting the pollutant diffusion trajectory according to the pollution source coordinates, correcting the pollutant diffusion trajectory by combining the traffic density data, and generating a pollutant heat map; superimposing the personnel distribution data with the pollutant heat map to obtain personnel avoidance paths, and determining the tangent direction in the personnel avoidance paths as the airflow direction of the evacuation guidance airflow, generating evacuation wind direction data; extracting the speed combination and deflection angle of the tunnel ventilation fan from the evacuation wind direction data, filtering the speed combination and deflection angle through a multi-objective reward mechanism, and outputting a fan control command set; adjusting the fan speed and deflection angle of the tunnel ventilation fan according to the fan control command set, and guiding directional airflow, and monitoring the pollutant concentration values ​​in each tunnel ventilation area after directional airflow guidance in real time through a smoke concentration sensor; comparing the predicted concentration values ​​in the pollutant heat map with the pollutant concentration values ​​to verify the tunnel ventilation control effect.

[0007] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the method involves: using a pre-trained pollutant diffusion model to predict the pollutant diffusion trajectory based on the pollution source coordinates, correcting the pollutant diffusion trajectory using traffic density data, and generating a pollutant heat map. The specific steps are as follows: Input the pollution source coordinates and wind speed and direction data in the tunnel into the pollutant diffusion model. The pollutant diffusion model uses the pollution source coordinates as the diffusion starting point and the wind speed and direction data in the tunnel as the guiding direction to simulate the pollutant diffusion and output the pollutant diffusion trajectory. Perform spatial gridding processing on the traffic flow density data to generate a traffic flow distribution matrix; The grid values ​​in the traffic flow distribution matrix are normalized, and the density diffusion effect is mapped onto the normalized grid values ​​to output the correction factor matrix. By spatially superimposing the correction factor matrix and the pollutant diffusion trajectory, the propagation speed and direction of pollutant diffusion are corrected, and a corrected pollutant diffusion trajectory is generated. The modified pollutant diffusion trajectory is converted into a two-dimensional gridded concentration distribution according to the time step, forming a pollutant heat map.

[0008] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the specific steps for overlaying personnel distribution data with a pollutant heat map to obtain personnel avoidance paths are as follows: Spatial gridding is performed on the personnel distribution data to generate a personnel distribution matrix; Spatially align the personnel distribution matrix with the pollutant heatmap to output a personnel-pollutant map; Traverse each grid in the personnel-pollutant map, compare the pollution concentration corresponding to each grid with the preset pollution threshold, and divide the safe zone and the risk zone; By treating high-risk areas as obstacles and safe areas as available paths, we can construct paths for personnel to avoid them.

[0009] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the specific steps for determining the tangential direction in the personnel avoidance path as the airflow direction for guiding evacuation airflow and generating evacuation wind direction data are as follows: Evacuation path points are extracted from the personnel avoidance paths, and the evacuation path points are sorted according to the order of the personnel avoidance paths to generate an ordered path chain. Extract the path vectors of the tangent direction for each pair of adjacent path points in the ordered path chain, and output the path vector sequence; The path vectors in the path vector sequence are converted into standard wind direction angles, and the standard wind direction angles are interpolated and smoothed to generate an airflow angle sequence. Spatial parameter mapping is performed between the airflow angle sequence and the pollutant heat map to generate evacuation wind direction data.

[0010] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the steps include: extracting the speed combination and deflection angle of the tunnel ventilation fan from the evacuation wind direction data, filtering the speed combination and deflection angle through a multi-objective reward mechanism, and outputting a fan control command set. Extract the standard wind direction angle from the evacuation wind direction data and record the grid coordinate range corresponding to the standard wind direction angle; Map the standard wind direction angles of each grid coordinate range to wind turbine parameters to obtain the speed combination and deflection angle; A multi-objective reward mechanism is used to score the rotation speed combination and deflection angle, and the rotation speed combination and deflection angle with the highest reward score are selected as candidate solutions for tunnel ventilation fans. The candidate solutions are converted into fan control commands and assigned to the corresponding tunnel ventilation fans according to the grid coordinate range, and the fan control command set is output.

[0011] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the step of scoring the rotation speed combination and deflection angle through a multi-objective reward mechanism refers to assigning reward weights to each rotation speed combination and deflection angle according to the energy consumption, evacuation efficiency and safety objectives of the tunnel ventilation fan, and calculating the sum of the reward weights of each rotation speed combination and deflection angle as the reward score.

[0012] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the steps include: adjusting the fan speed and deflection angle of the tunnel ventilation fan according to the fan control command set, guiding directional airflow, and monitoring the pollutant concentration values ​​in each tunnel ventilation area in real time using a smoke concentration sensor. The specific steps are as follows: The fan control command set is converted into PLC communication protocol signals and sent to the fan drive center of the corresponding tunnel ventilation fan; The fan drive center adjusts the fan speed and deflection angle of the tunnel ventilation fan to create directional airflow guidance; After directional airflow guidance, the pollutant concentration values ​​of each tunnel ventilation area are obtained by smoke concentration sensors at fixed time intervals.

[0013] As a preferred embodiment of the tunnel ventilation control method based on environmental prediction described in this invention, the specific steps for comparing the predicted concentration values ​​in the pollutant heat map with the actual pollutant concentration values ​​to verify the tunnel ventilation control effect are as follows: Extract the predicted concentration value of each grid in the pollutant heatmap and record the grid coordinate range corresponding to the predicted concentration value; The predicted concentration values ​​and pollutant concentration values ​​are spatiotemporally aligned according to the grid coordinate range to generate pollutant data pairs. The predicted concentration value and the actual concentration value of pollutants in the data pair are compared point by point, the pollutant concentration difference is output, and a concentration threshold is set. If the pollutant concentration difference is positive and the absolute value of the pollutant concentration difference does not exceed the concentration threshold, then the ventilation control effect is deemed to be up to standard. If the pollutant concentration difference is negative, the grid coordinate range corresponding to the pollutant concentration value is marked as an abnormal area, and an early warning iteration mechanism is triggered.

[0014] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the tunnel ventilation control method based on environmental prediction as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the tunnel ventilation control method based on environmental prediction as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: By synergistically generating pollutant heat maps and evacuation direction data, environmental prediction and airflow guidance correction are achieved during tunnel ventilation control. By combining pollutant source coordinates, tunnel wind speed and direction data, and traffic density data to correct pollutant diffusion trajectories, a pollutant heat map with spatial distribution characteristics is formed. This provides continuous concentration data for constructing personnel avoidance paths. The tangential direction in the personnel avoidance paths is determined as the airflow direction for evacuation guidance, generating evacuation direction data. This provides a clear directional basis for the formation of the fan control command set, thereby supporting the verification of directional airflow guidance and ventilation control effects, and improving the directionality, consistency, and targeted execution of tunnel ventilation control. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a tunnel ventilation control method based on environmental prediction.

[0019] Figure 2 This is a flowchart for generating a pollutant heat map.

[0020] Figure 3 A flowchart for generating evacuation wind direction data.

[0021] Figure 4 A flowchart for verifying the effectiveness of ventilation control. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a tunnel ventilation control method based on environmental prediction, including the following steps: S1. Collect environmental data and perform preprocessing.

[0026] Environmental data includes pollution source coordinates, traffic density data, wind speed and direction data inside the tunnel, and personnel distribution data; Specifically, pollution source coordinates are collected using a fixed gas sensor array. These fixed gas sensors are deployed at tunnel entrances and ventilation nodes to record the three-dimensional spatial coordinates (x, y, z) and initial diffusion time of the pollution source. Wind speed and direction data within the tunnel are collected using distributed anemometers and wind vanes. The distributed anemometers are installed on the tunnel top and sidewalls, while the wind vanes acquire wind direction information synchronously through laser ranging and angle encoders. Traffic density data is collected jointly using a geomagnetic sensor and a video recognition device. The geomagnetic sensor is buried in the road surface to monitor the frequency and speed of passing vehicles in real time. The video recognition device extracts vehicle outlines and counts the number of vehicles per unit area using image segmentation algorithms. Personnel distribution data is collected jointly using a thermal imaging camera and mobile terminal signal positioning. The thermal imaging camera is deployed at tunnel exits and intersections to capture the distribution of human body heat radiation, while mobile terminal signal positioning obtains personnel location coordinates through Wi-Fi probes and Bluetooth beacons.

[0027] Preprocessing includes data cleaning, timestamp alignment, and spatial coordinate calibration; Specifically, data cleaning addresses outliers and missing values ​​in pollution source coordinates, wind speed and direction data, traffic density data, and personnel distribution data. Data points exceeding the tunnel's physical boundaries in pollution source coordinates are removed, and missing coordinates are supplemented using nearest neighbor interpolation. For discrete noise in wind speed and direction data caused by equipment malfunctions, a sliding window averaging method is used to smooth abnormal fluctuations; for continuous missing values, linear interpolation is used to restore data continuity. For abrupt changes in traffic density data caused by occlusion or identification errors, a moving median filter based on time series is used to remove abnormal peaks. For isolated points in personnel distribution data caused by signal interference, a clustering algorithm based on spatial neighborhood density is used to identify and remove isolated points. When identifying and removing isolated points, all coordinate points in the personnel distribution data are first mapped to the tunnel's three-dimensional spatial grid. A local neighborhood is set with each coordinate point as the center. The number of adjacent coordinate points within the local neighborhood is counted as a local density index. If the number of adjacent coordinate points within the local neighborhood of a coordinate point is lower than a preset coordinate point threshold (the preset coordinate point threshold is dynamically set by statistically analyzing the local neighborhood density distribution characteristics of the personnel distribution data under normal conditions), the coordinate point is determined to be an isolated point in a low-density area. This indicates that there is a lack of effective signal support around the coordinate point, which may be caused by signal interference or positioning drift. The confirmed isolated points are removed from the dataset, while the reliable personnel location information in high-density clustered areas is retained, thereby improving the spatial accuracy of the personnel distribution data. Timestamp alignment unifies the time fields of pollution source coordinates, wind speed and direction data, traffic density data, and personnel distribution data to the same time base. First, the original timestamps of each environmental data are extracted, and differences in timestamp format (such as ISO 8601, Unix timestamps, etc.) are identified. All data are converted to millisecond-level timestamps in the UTC+8 time zone. Then, the timestamps of environmental data (such as pollution source coordinates) are selected as the base, and the timestamps are aligned to the time interval of the pollution source coordinates (such as once every 10 seconds) using linear interpolation to ensure the temporal consistency between pollution diffusion simulation and dynamic environmental data. Spatial coordinate calibration unifies the spatial coordinates of pollution source coordinates, wind speed and direction data, vehicle flow density data, and personnel distribution data into the tunnel's three-dimensional coordinate system. First, a global coordinate system for the tunnel is established, with the tunnel entrance centerline as the positive x-axis, the tunnel longitudinal centerline as the positive y-axis, and ground height as the zero point of the z-axis. The latitude and longitude or relative coordinates of the pollution source coordinates are projected and transformed to ensure alignment with the tunnel coordinate system. The angle information of the wind vane in the wind speed and direction data is converted into azimuth angles relative to the tunnel coordinate system (e.g., 0° represents the positive x-axis, and 90° represents the positive y-axis). The vehicle position coordinates in the vehicle flow density data are mapped to the tunnel coordinate system through perspective transformation. The coordinates of personnel distribution data located by thermal imaging and motion signals are superimposed to eliminate spatial deviations of multi-source data, ultimately generating environmental data under a unified coordinate system.

[0028] S2. Based on the pre-trained pollutant diffusion model, the pollutant diffusion trajectory is predicted according to the pollution source coordinates, and the pollutant diffusion trajectory is corrected by combining traffic flow density data to generate a pollutant heat map.

[0029] Input the pollution source coordinates and wind speed and direction data in the tunnel into the pollutant diffusion model. The pollutant diffusion model uses the pollution source coordinates as the diffusion starting point and the wind speed and direction data in the tunnel as the guiding direction to simulate the pollutant diffusion and output the pollutant diffusion trajectory. Specifically, the pollutant diffusion model is jointly constructed based on the physical diffusion equation and machine learning algorithms. First, a framework for the pollutant diffusion model is established based on the convection-diffusion equation. The tunnel space is divided into grids (e.g., dividing the tunnel into fixed-size three-dimensional grids). The pollution source coordinates (x, y, z), initial diffusion time, and wind speed and direction data (wind speed value and wind direction angle) are mapped to grid nodes to obtain the change in pollutant concentration over time, generating a basic pollutant diffusion trajectory. Then, historical environmental data (such as pollution source coordinates, wind speed and direction data, and measured pollutant concentration distribution) is used as the training set. Supervised learning is used to optimize the parameters of the pollutant diffusion model. Specifically, pollution source coordinates, wind speed and direction data, and traffic density data are used as input features, along with measured pollutant concentration data. The data is labeled with a distribution, and a convolutional neural network or graph neural network is used to process the gridded spatial data. Physical equations (such as convection-diffusion equations) are embedded as constraints to generate a pre-trained pollutant diffusion model. The pollutant diffusion model takes the pollution source coordinates as the diffusion starting point and combines the wind speed and wind direction data to simulate the diffusion path of pollutants in the three-dimensional space of the tunnel. The pollutant diffusion model divides the tunnel space into grids and maps the wind speed and wind direction data to grid nodes. It obtains the change of pollutant concentration over time for each grid node and outputs the pollutant diffusion trajectory as a series of pollutant concentration distribution layers corresponding to a series of time steps. Each grid in the layer records the predicted pollutant concentration value and its position coordinates in space, and finally generates a dynamic pollutant diffusion trajectory sequence.

[0030] Perform spatial gridding processing on the traffic flow density data to generate a traffic flow distribution matrix; Specifically, traffic density data is the spatiotemporal distribution of vehicle frequency and speed. First, the tunnel space is divided into a fixed-size two-dimensional grid (e.g., 1 meter × 1 meter). The number of vehicles in each grid is counted to generate a traffic distribution matrix. Each element in the traffic distribution matrix corresponds to the traffic density value of each grid. The traffic density value is obtained by counting the number of vehicles per unit area. The number of rows and columns of the traffic distribution matrix corresponds to the number of grid divisions in the longitudinal and transverse directions of the tunnel, respectively. The value range of the matrix is ​​dynamically adjusted according to the actual traffic flow and is used for the generation and superposition of subsequent correction factors. The grid values ​​in the traffic flow distribution matrix are normalized, and the density diffusion effect is mapped onto the normalized grid values ​​to output the correction factor matrix. Each grid value in the traffic flow distribution matrix is ​​transformed using the min-max normalization method. After normalization, the normalized value is mapped to a correction factor based on the influence of traffic flow density on pollutant diffusion. The mapping rule for the correction factor is as follows: the higher the traffic flow density, the larger the correction factor (e.g., 0.8-1.0), indicating the enhancing effect of vehicle exhaust emissions on pollutant diffusion; the lower the traffic flow density, the smaller the correction factor (e.g., 0.2-0.4), indicating the suppressing effect of airflow disturbance on pollutant diffusion. Finally, a correction factor matrix is ​​generated, and the correction factor value of each element in the correction factor matrix corresponding to the grid is used for subsequent correction of pollutant diffusion trajectories.

[0031] By spatially superimposing the correction factor matrix and the pollutant diffusion trajectory, the propagation speed and direction of pollutant diffusion are corrected, and a corrected pollutant diffusion trajectory is generated. The correction factor matrix and the pollutant diffusion trajectory are superimposed grid by grid in the spatial dimension. The specific process of grid-by-grid superposition is as follows: the pollutant concentration value of each grid in the pollutant diffusion trajectory is weighted with the correction factor value of the corresponding grid in the correction factor matrix to adjust the propagation intensity of the pollutant concentration. At the same time, the wind speed value in the wind speed and wind direction data is adjusted according to the numerical ratio of the correction factor, while the wind direction angle is kept consistent with the original wind speed and wind direction data. By combining the superimposed pollutant concentration value and the corrected wind speed value with the spatial distribution of wind speed and wind direction data in the tunnel, the diffusion path of pollutants in the three-dimensional space of the tunnel is obtained, and the corrected pollutant diffusion trajectory is generated. The pollutant concentration value and propagation direction of each grid in the corrected pollutant diffusion trajectory reflect the influence of traffic density on the diffusion process. The modified pollutant diffusion trajectory is converted into a two-dimensional gridded concentration distribution according to the time step, forming a pollutant heat map; The process corrects the pollutant diffusion trajectory by including multiple time steps (e.g., every 10 seconds) of pollutant concentration distribution layers, with each layer corresponding to the spatial distribution of pollutant concentration at a given time point. The pollutant concentration distribution layers corresponding to each time step are then converted into a two-dimensional gridded concentration matrix. The number of rows and columns in the two-dimensional gridded concentration matrix corresponds to the number of grid divisions in the tunnel's longitudinal and transverse directions. The numerical range of the two-dimensional gridded concentration matrix is ​​dynamically adjusted based on the pollutant concentration. Finally, the two-dimensional gridded concentration matrix is ​​visualized using color gradients to generate a pollutant heatmap. High-concentration areas in the pollutant heatmap are represented by red or orange, while low-concentration areas are represented by green or blue, visually demonstrating the diffusion range and intensity of pollutants in the tunnel space.

[0032] By jointly constructing pollutant diffusion trajectories using physical diffusion equations and machine learning algorithms, and dynamically adjusting the diffusion path by incorporating traffic density correction factors, the prediction accuracy and environmental adaptability are significantly improved. By embedding convection-diffusion equations as constraints, the pollutant diffusion model follows physical laws while possessing data-driven flexibility, reducing prediction errors. Furthermore, the introduction of traffic density correction factors compensates for the traditional model's neglect of traffic disturbances, enabling the pollutant diffusion trajectory to respond to traffic changes in real time and avoid missing high-pollution areas. This not only improves the adaptability of the pollutant diffusion model to dynamic environments but also optimizes parameters using historical data, reducing abnormal deviations caused by environmental noise or measurement errors, and providing a reliable basis for subsequent early warning mechanisms.

[0033] S3. Overlay the personnel distribution data with the pollutant heat map to obtain the personnel avoidance path, and determine the tangent direction in the personnel avoidance path as the airflow direction of the evacuation guidance airflow to generate evacuation wind direction data.

[0034] Spatial gridding is performed on the personnel distribution data to generate a personnel distribution matrix; Specifically, the personnel distribution data is a set of coordinates of human body thermal radiation distribution and mobile terminal signal positioning. First, the tunnel space is divided into a fixed-size two-dimensional grid (such as 1 meter × 1 meter), and the number of people in each grid is counted to generate a personnel distribution matrix. Each element in the personnel distribution matrix corresponds to the personnel density value of the grid. The personnel density value is obtained by counting the number of people per unit area. The number of rows and columns of the personnel distribution matrix correspond to the number of grid divisions in the longitudinal and transverse directions of the tunnel, respectively. Spatially align the personnel distribution matrix with the pollutant heatmap to output a personnel-pollutant map; Both the personnel distribution matrix and the pollutant heat map are based on a unified spatial division of the tunnel's three-dimensional coordinate system. They are aligned in spatial dimensions through coordinate mapping. Specifically, the coordinates of each grid in the personnel distribution matrix are matched with the corresponding grid coordinates in the pollutant heat map to ensure that the grid divisions on the x-axis (positive direction of the tunnel entrance centerline) and y-axis (positive direction of the tunnel longitudinal centerline) are consistent. The aligned datasets are combined to form a personnel-pollutant map. Each grid in the personnel-pollutant map contains the personnel density value and the corresponding pollutant concentration value, forming a spatial correlation feature between personnel and pollution concentration. Traverse each grid in the personnel-pollutant map, compare the pollution concentration corresponding to each grid with the preset pollution threshold, and divide the safe zone and the risk zone; The setting of preset pollution thresholds is typically determined based on industry requirements and historical environmental data statistical analysis. The specific process includes: first, referring to the pollutant concentration limits stipulated in relevant regulations, combined with the acute or chronic effects of pollutants on human health; second, determining a reasonable range for the preset pollution threshold by statistically analyzing the normal fluctuation range of pollutants through long-term monitoring data; and finally, dynamically adjusting based on actual environmental characteristics (such as tunnel ventilation conditions, personnel exposure time, etc.) to ensure that the preset pollution threshold both guarantees safety and avoids oversensitivity leading to misjudgment. The preset pollution threshold is the safe upper limit of pollutant concentration (e.g., 0.5 mg / m³). By traversing the personnel-pollutant map grid by grid, the pollution concentration of each grid is compared with the preset pollution threshold. If the pollutant concentration value is less than or equal to the preset threshold, the grid is marked as a safe area; if the pollutant concentration value is greater than the preset threshold, the grid is marked as a risk area. The division of safe and risk areas forms a binary spatial layer, used for subsequent construction of personnel avoidance paths.

[0035] By treating high-risk areas as obstacles and safe areas as available paths, we can construct paths for personnel to avoid them. Specifically, based on the binary spatial layers of safe and risk areas, path planning algorithms (such as A* or Dijkstra's algorithm) are used to construct personnel avoidance paths; risk areas are treated as impassable obstacles, limiting the path search range to only within safe areas; the path planning algorithm uses the grid cells of the personnel density area in the personnel distribution matrix as the starting point and the tunnel exit or safe refuge area as the ending point to obtain the shortest feasible path from the starting point to the ending point; during the path search process, paths with low personnel density but high connectivity in safe areas are prioritized to generate personnel avoidance paths; Evacuation path points are extracted from the personnel avoidance paths, and the evacuation path points are sorted according to the order of the personnel avoidance paths to generate an ordered path chain. The personnel avoidance path consists of a series of continuous grids, each grid corresponding to spatial coordinates. The coordinates of all grids in the path are extracted as evacuation path points. The evacuation path points are sorted according to the traversal order of the path search algorithm (such as the stepping direction from the starting point to the end point) to generate an ordered path chain. Each path point in the ordered path chain records the x and y coordinates in the tunnel space and the path sequence number, which are used for subsequent path vector extraction and wind direction angle acquisition. Extract the path vectors of the tangent direction for each pair of adjacent path points in the ordered path chain, and output the path vector sequence; In an ordered path chain, adjacent evacuation path points form continuous path segments. The tangent direction of each path segment is determined by the coordinate difference between adjacent path points. The specific process is as follows: Calculate the x-axis coordinate difference and y-axis coordinate difference between each pair of adjacent path points to form a path vector (Δx, Δy). The direction of the path vector is determined by the ratio of Δx to Δy. For example, Δx=1, Δy=0 indicates the positive x-axis direction, and Δx=0, Δy=1 indicates the positive y-axis direction. Arrange the path vectors of all path segments in order to generate a path vector sequence. The path vectors in the path vector sequence are converted into standard wind direction angles, and the standard wind direction angles are interpolated and smoothed to generate an airflow angle sequence. Each path vector (Δx, Δy) in the path vector sequence is used to calculate the standard wind direction angle using the arctangent function. The transformed standard wind direction angle is set to 0° as the positive x-axis and 90° as the positive y-axis to ensure consistency with the azimuth definition of the tunnel coordinate system. The standard wind direction angles are sorted according to the sequence order of the path vector sequence to generate a wind direction angle sequence. The generated wind direction angle sequence is then interpolated (e.g., linear interpolation or spline interpolation) to fill the angle gaps between adjacent vectors in the path vector sequence. Subsequently, the interpolated wind direction angle sequence is smoothed (e.g., using the sliding window averaging method) to eliminate abrupt angle changes caused by path turns, generating a continuous and smooth airflow angle sequence. The standard wind direction angle formula is: ; in, This represents the standard wind direction angle for each path vector. This represents the difference in x-axis coordinates between adjacent path points. This represents the difference in y-axis coordinates between adjacent path points. Represents the identifier for pi; Spatial parameter mapping is performed between airflow angle sequences and pollutant heat maps to generate evacuation wind direction data; Each angle value in the airflow angle sequence corresponds to the spatial location of an evacuation path point in an ordered path chain. The standard wind direction angle is associated with the grid in the pollutant heat map through spatial parameter mapping. Specifically, the x and y coordinates of each path point are aligned with the grid division of the pollutant heat map to determine the grid position of the path point in the heat map. The corresponding standard wind direction angle is written into the wind direction parameter field of the grid to form evacuation wind direction data. The evacuation wind direction data is stored in the form of a two-dimensional matrix.

[0036] By overlaying personnel distribution data with pollutant heat maps to generate avoidance paths and converting the path tangent direction into airflow angles, a two-way adaptation between personnel evacuation and airflow guidance is achieved, significantly enhancing evacuation safety and efficiency. The path planning method, which overlays personnel distribution data with pollutant heat maps to generate avoidance paths, ensures that personnel avoid highly polluted areas. At the same time, through the smooth mapping of path vector sequences and wind direction angles, the airflow distribution continuously covers the evacuation path, shortening the time for pollutant concentration to decrease (e.g., increasing the concentration decrease rate in highly polluted areas by 30%). In addition, the dynamic adaptation capability can adjust the airflow strategy in real time when personnel density changes, adapting to the evacuation needs in the event of sudden passenger flow or accident scenarios. Thus, while ensuring the safe evacuation of personnel, it efficiently dilutes pollutants and reduces health risks.

[0037] S4. Extract the speed combination and deflection angle of the tunnel ventilation fan from the evacuation wind direction data, filter the speed combination and deflection angle through a multi-objective reward mechanism, and output the fan control command set.

[0038] Extract the standard wind direction angle from the evacuation wind direction data and record the grid coordinate range corresponding to the standard wind direction angle; Specifically, the evacuation wind direction data is stored in the form of a two-dimensional matrix. Each grid in the two-dimensional matrix contains the standard wind direction angle value and the corresponding spatial coordinates. By traversing all the grids in the evacuation wind direction data, the standard wind direction angle of each grid is extracted, and the grid coordinate range corresponding to the standard wind direction angle is recorded. The grid coordinate range is represented in the form of continuous intervals or discrete points to ensure the accuracy of subsequent wind turbine parameter mapping. Map the standard wind direction angles of each grid coordinate range to wind turbine parameters to obtain the speed combination and deflection angle; The fan parameter mapping is based on the physical layout and functional characteristics of the tunnel ventilation fans. First, the corresponding tunnel ventilation fan positions are determined according to the grid coordinate range. Then, the standard wind direction angle is converted into the fan speed combination and deflection angle. The speed combination is determined by the relationship table between wind speed and air volume, and the deflection angle is determined by the mapping table between wind direction angle and fan blade rotation angle. For example, if the standard wind direction angle is 45°, the fan blades need to be adjusted to the 45° direction, and the corresponding speed is matched according to the wind speed requirement. During the fan parameter mapping process, it is necessary to combine the correction of pollutant diffusion trajectory to ensure that the speed combination and deflection angle can effectively cover the high-pollution area and guide the airflow to diffuse to the safe area. Furthermore, the wind speed and air volume relationship table is generated based on the wind speed data obtained during the environmental data acquisition phase and the results of the tunnel spatial grid division. After preprocessing, the wind speed values ​​collected by distributed anemometers and wind vanes are combined with the grid divided in the tunnel three-dimensional coordinate system. Using the performance curves provided by the wind turbine manufacturer as a reference, the actual measured wind speed is correlated with the air volume during wind turbine operation. By monitoring the wind turbine operating status at different speeds over a long period of time, and combining the physical correlation between wind speed and air volume, a dynamic mapping relationship between speed combination and air volume is established to ensure that the wind turbine control strategy can accurately match the airflow demand. The mapping table between wind direction angle and wind turbine blade rotation angle relies on wind direction data collected by the wind vane and the physical layout characteristics of the wind turbine. The wind vane obtains wind direction information synchronously through laser ranging and angle encoder. During the spatial coordinate calibration stage, it has been converted into the azimuth angle definition of the tunnel coordinate system, that is, 0° represents the positive x-axis direction and 90° represents the positive y-axis direction. The rotation angle of the wind turbine blade must strictly correspond to the azimuth angle. At the same time, it is calibrated and corrected by combining the rotation direction difference (such as right-hand or left-hand rotation) during wind turbine installation. By matching the actual records of wind direction changes and wind turbine blade adjustments in historical environmental data, a one-to-one mapping relationship between wind direction angle and blade rotation angle is established to ensure that the airflow guidance direction is consistent with the evacuation path design.

[0039] A multi-objective reward mechanism is used to score the rotation speed combination and deflection angle, and the rotation speed combination and deflection angle with the highest reward score are selected as candidate solutions for tunnel ventilation fans. Specifically, the multi-objective reward mechanism for evaluating speed combinations and deflection angles involves assigning reward weights to each speed combination and deflection angle based on the tunnel ventilation fan's energy consumption, evacuation efficiency, and safety objectives. The sum of these reward weights is then used as the reward score. First, energy consumption is quantified by the product of fan speed and operating time; higher speeds and longer operating times result in higher energy consumption, and vice versa. Evacuation efficiency is measured by the rate of pollutant concentration reduction, such as the concentration decrease in high-pollution areas within a preset time and the expansion area of ​​safe zones. Safety objectives are assessed through the coverage of personnel avoidance paths to ensure airflow. The direction of the tunnel ventilation fan highly overlaps with the personnel evacuation route to avoid airflow disturbances that could lead to personnel being stranded or subjected to secondary exposure. The reward weights for energy consumption, evacuation efficiency, and safety objectives are optimized based on historical environmental data. For example, the reward weights are increased for improving evacuation efficiency in high-pollution scenarios and for improving energy consumption efficiency during periods of energy shortage. The reward score for each combination of rotation speed and deflection angle is generated by weighting three parts: a negative score for energy consumption (the lower the energy consumption, the higher the score), a positive score for evacuation efficiency (the better the evacuation effect, the higher the score), and a positive score for safety objectives (the wider the path coverage, the higher the score). Finally, the scheme with the highest comprehensive score is selected as the candidate scheme for the tunnel ventilation fan. The candidate solutions are converted into fan control commands and assigned to the corresponding tunnel ventilation fans according to the grid coordinate range, and the fan control command set is output. The speed percentage value (e.g., 80%) in the candidate scheme is quantized and converted according to the variable frequency speed control protocol supported by the wind turbine drive center (e.g., 0-10V analog signal or Modbus RTU protocol). For example, 80% speed corresponds to a voltage value of 6.4V or a hexadecimal register value of 0x0640. Simultaneously, the deflection angle value is converted into a servo motor control command based on the resolution of the wind turbine blade rotation (e.g., 1 pulse signal per 0.1°). Then, the quantized speed and deflection angle parameters are encapsulated according to the standard frame format of the PLC communication protocol. For example, ASCII or binary format is used to define the instruction start character, data fields (speed percentage and deflection angle), check bits (e.g., CRC16 checksum), and end character to ensure the integrity and anti-interference of instruction transmission. The control commands include specific values ​​for speed combinations and deflection angles. The format of the control commands must be compatible with the communication protocol of the tunnel ventilation fans to ensure the accuracy of the executed commands. When allocating control commands according to the grid coordinate range, multiple fan control commands within the same grid coordinate range must be merged into a fan control command set to avoid redundant operations. The final output fan control command set is organized according to time step and spatial location, and the evacuation airflow direction data is kept synchronized to ensure dynamic adaptation of ventilation strategies and personnel avoidance paths. The fan control command set is sent to each tunnel ventilation fan through the tunnel ventilation execution center to complete the real-time control of the airflow direction and intensity within the tunnel.

[0040] The wind turbine control commands are optimized based on a multi-objective reward mechanism, balancing energy consumption, evacuation efficiency, and safety objectives. This addresses the suboptimal solutions caused by traditional single-indicator optimization. By comprehensively evaluating energy consumption inverse scoring (high score for low energy consumption), evacuation efficiency positive scoring (rapid pollution reduction), and safety scoring (path coverage), it avoids over-reliance on any single indicator. For example, the weight of evacuation efficiency is dynamically increased in high-pollution scenarios, or energy consumption is prioritized to be reduced during periods of energy shortage, ensuring that the wind turbine strategy always aligns with actual needs.

[0041] S5. Adjust the fan speed and deflection angle of the tunnel ventilation fan according to the fan control command set, and guide the directional airflow. Monitor the pollutant concentration value of each tunnel ventilation area in real time after the directional airflow is guided by the smoke concentration sensor.

[0042] The fan control command set is converted into PLC communication protocol signals and sent to the fan drive center of the corresponding tunnel ventilation fan; Specifically, firstly, the speed combinations and deflection angle values ​​in the fan control instruction set are mapped to register addresses or signal channels in the PLC communication protocol. For example, the speed percentage value is converted into 16-bit integer data, and the deflection angle value is converted into floating-point data from 0 to 360°. Then, the PLC communication protocol signal is sent to the fan drive center of the tunnel ventilation fan through serial communication, Ethernet, or wireless transmission. During the transmission process, a check code (such as CRC check) needs to be added to ensure data integrity and avoid incorrect adjustment of fan parameters due to transmission errors. The fan drive center adjusts the fan speed and deflection angle of the tunnel ventilation fan to create directional airflow guidance; The fan drive center controls the frequency converter to adjust the fan speed to the target percentage value according to the received PLC communication protocol signal. At the same time, it drives the servo motor to rotate the blades to the specified deflection angle to ensure that the airflow direction is consistent with the evacuation path design. The adjustment process needs to combine the mapping logic of speed combination and deflection angle. For example, if the target deflection angle is 45°, the servo motor needs to rotate to an azimuth angle of 45° with the positive x-axis of the tunnel coordinate system. At the same time, the corresponding speed is matched according to the relationship table of wind speed and air volume to ensure that the airflow intensity covers the high-pollution area and promotes the diffusion of pollutants to the safe area. After the directional airflow guidance is completed, the airflow direction and intensity in the tunnel need to be synchronized with the evacuation wind direction data to ensure the continuity of airflow coverage of personnel avoidance path. After directional airflow guidance, the pollutant concentration values ​​of each tunnel ventilation area are obtained by smoke concentration sensors at fixed time intervals; Smoke concentration sensors are deployed in the ventilation areas of the tunnel, including near pollution sources, high-density traffic areas, and along pedestrian avoidance paths. The smoke concentration sensors detect pollutant concentrations based on electrochemical or optical principles. According to the spatial grid division rules of the pollutant heat map, the smoke concentration sensors periodically sample the pollutant concentration in each grid area. The sampling frequency is synchronized with the time step of the fan control command set (e.g., once every 10 seconds) to ensure that the monitoring data matches the airflow guidance strategy in real time. The sampled data is uploaded to the data processing center via wired or wireless network. After the data processing center performs a preprocessing process on the pollutant concentration values ​​(including data cleaning, timestamp alignment, and spatial coordinate calibration), a standardized pollutant concentration distribution layer is generated.

[0043] S6. Compare the predicted concentration values ​​in the pollutant heat map with the pollutant concentration values ​​to verify the effectiveness of tunnel ventilation control.

[0044] Extract the predicted concentration value of each grid in the pollutant heatmap and record the grid coordinate range corresponding to the predicted concentration value; Specifically, the pollutant heatmap contains a two-dimensional gridded concentration matrix with multiple time steps (e.g., once every 10 seconds). Each grid stores the predicted pollutant concentration value and its corresponding spatial coordinates. The extraction process involves traversing each grid in the heatmap to obtain the predicted concentration value (e.g., 0.4 mg / m³) and the x-axis coordinate (e.g., x=15m) and y-axis coordinate (e.g., y=8m), and recording the coordinate range of the grid (e.g., x∈[10m, 20m], y∈[5m, 10m]). The predicted concentration values ​​and pollutant concentration values ​​are spatiotemporally aligned according to the grid coordinate range to generate pollutant data pairs. The spatiotemporal alignment process requires matching the time step of the predicted concentration value (e.g., t=10 seconds) with the sampling time of the actual monitoring data (e.g., t=10 seconds) to ensure consistency in the time dimension; at the same time, the grid coordinates of the predicted concentration value are aligned with the grid coordinates of the actual monitoring data to ensure consistency in the spatial dimension; after alignment, the predicted concentration value and the actual monitoring concentration value of the same grid at the same time step are included to form a pollutant data pair (predicted value, measured value).

[0045] The predicted concentration value and the actual concentration value of pollutants in the data pair are compared point by point, the pollutant concentration difference is output, and a concentration threshold is set. The difference between the predicted concentration value and the actual concentration value in the pollutant data pair is calculated by subtracting them point by point. The concentration threshold is set based on industry requirements and historical environmental data to obtain the mean, standard deviation or confidence interval of the pollutant concentration and determine the range of the pollutant concentration difference (e.g., ±0.05 mg / m³). The concentration threshold must cover all fluctuation ranges to avoid misjudgment due to environmental noise or measurement errors, and at the same time ensure that an early warning mechanism is triggered when the deviation exceeds the threshold. The formula for calculating the pollutant concentration difference is: ; in, This represents the difference between the predicted and measured concentrations of pollutants. This represents the predicted concentration value in the pollutant data pair. This represents the pollutant concentration value in the pollutant data pair; If the pollutant concentration difference is positive and the absolute value of the pollutant concentration difference does not exceed the concentration threshold, then the ventilation control effect is deemed to be up to standard. If the pollutant concentration difference is negative, the grid coordinate range corresponding to the pollutant concentration value is marked as an abnormal area, and an early warning iteration mechanism is triggered. When the pollutant concentration difference is negative, it indicates that the pollutant diffusion model underestimates the actual pollutant concentration, which may indicate insufficient airflow guidance or that the pollutant diffusion path does not completely cover the high-pollution area. The corresponding grid coordinate range (e.g., x∈[10m, 20m], y∈[5m, 10m]) needs to be marked as an abnormal area, and an early warning iteration mechanism should be triggered: First, retrieve the weight allocation parameters of the multi-objective reward mechanism and re-evaluate the priority of energy consumption, evacuation efficiency, and safety objectives; Second, combine the real-time operating data of the fan drive center (e.g., speed percentage, blade deflection angle) to adjust the fan control command set and generate new speed combinations and deflection angle candidate schemes; Finally, screen the optimized scheme through the multi-objective reward scoring process and send the updated command set to the fan drive center to achieve dynamic correction of the ventilation strategy.

[0046] By comparing pollutant diffusion with pollutant concentration values, a prediction-verification closed loop is formed to ensure continuous optimization of ventilation control effects. By combining point-by-point pollutant concentration difference analysis to distinguish between normal fluctuations and abnormal deviations, misjudgments are effectively avoided. This not only improves the adaptability of the pollutant diffusion model, but also enables continuous iteration of ventilation strategies through dynamic correction, providing stable technical support for pollutant control and personnel evacuation in complex scenarios.

[0047] This embodiment also provides a computer device applicable to the tunnel ventilation control method based on environmental prediction, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the tunnel ventilation control method based on environmental prediction as proposed in the above embodiment.

[0048] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0049] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the tunnel ventilation control method based on environmental prediction as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0050] In summary, this invention achieves environmental prediction and airflow guidance correction during tunnel ventilation control through the synergy of pollutant heat map generation and evacuation direction data generation. By combining pollutant source coordinates, tunnel wind speed and direction data, and traffic density data to correct pollutant diffusion trajectories, a pollutant heat map with spatial distribution characteristics is formed. This provides continuous concentration data for constructing personnel avoidance paths. The tangential direction in the personnel avoidance paths is determined as the airflow direction for evacuation guidance airflow, generating evacuation direction data. This provides a clear directional basis for the formation of the fan control command set, thereby supporting the verification of directional airflow guidance and ventilation control effects, and improving the directionality, consistency, and targeted execution of tunnel ventilation control.

[0051] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A tunnel ventilation control method based on environmental prediction, characterized in that: include, Collect environmental data and preprocess it; the environmental data includes pollution source coordinates, traffic density data, wind speed and direction data inside the tunnel, and personnel distribution data; Based on a pre-trained pollutant diffusion model, the pollutant diffusion trajectory is predicted according to the coordinates of the pollution source. The pollutant diffusion trajectory is then corrected by combining traffic density data to generate a pollutant heat map. By overlaying personnel distribution data with pollutant heat maps, personnel avoidance paths are obtained, and the tangent direction in the personnel avoidance paths is determined as the airflow direction of the evacuation guidance airflow, generating evacuation wind direction data. Extract the speed combination and deflection angle of the tunnel ventilation fan from the evacuation wind direction data, filter the speed combination and deflection angle through a multi-objective reward mechanism, and output the fan control command set; The fan speed and deflection angle of the tunnel ventilation fan are adjusted according to the fan control command set, and directional airflow is guided. The pollutant concentration value of each tunnel ventilation area is monitored in real time by the smoke concentration sensor after the directional airflow is guided. The predicted concentration values ​​in the pollutant heat map are compared with the actual pollutant concentration values ​​to verify the effectiveness of tunnel ventilation control.

2. The tunnel ventilation control method based on environmental prediction as described in claim 1, characterized in that: The pre-trained pollutant diffusion model predicts the pollutant diffusion trajectory based on the pollution source coordinates, corrects the trajectory by incorporating traffic density data, and generates a pollutant heat map. The specific steps are as follows: Input the pollution source coordinates and wind speed and direction data in the tunnel into the pollutant diffusion model. The pollutant diffusion model uses the pollution source coordinates as the diffusion starting point and the wind speed and direction data in the tunnel as the guiding direction to simulate the pollutant diffusion and output the pollutant diffusion trajectory. Perform spatial gridding processing on the traffic flow density data to generate a traffic flow distribution matrix; The grid values ​​in the traffic flow distribution matrix are normalized, and the density diffusion effect is mapped onto the normalized grid values ​​to output the correction factor matrix. By spatially superimposing the correction factor matrix and the pollutant diffusion trajectory, the propagation speed and direction of pollutant diffusion are corrected, and a corrected pollutant diffusion trajectory is generated. The modified pollutant diffusion trajectory is converted into a two-dimensional gridded concentration distribution according to the time step, forming a pollutant heat map.

3. The tunnel ventilation control method based on environmental prediction as described in claim 2, characterized in that: The specific steps for overlaying personnel distribution data with pollutant heat maps to obtain personnel avoidance paths are as follows: Spatial gridding is performed on the personnel distribution data to generate a personnel distribution matrix; Spatially align the personnel distribution matrix with the pollutant heatmap to output a personnel-pollutant map; Traverse each grid in the personnel-pollutant map, compare the pollution concentration corresponding to each grid with the preset pollution threshold, and divide the safe zone and the risk zone; By treating high-risk areas as obstacles and safe areas as available paths, we can construct paths for personnel to avoid them.

4. The tunnel ventilation control method based on environmental prediction as described in claim 3, characterized in that: The specific steps for determining the tangential direction in the personnel avoidance path as the airflow direction for guiding evacuation airflow and generating evacuation wind direction data are as follows: Evacuation path points are extracted from the personnel avoidance paths, and the evacuation path points are sorted according to the order of the personnel avoidance paths to generate an ordered path chain. Extract the path vectors of the tangent direction for each pair of adjacent path points in the ordered path chain, and output the path vector sequence; The path vectors in the path vector sequence are converted into standard wind direction angles, and the standard wind direction angles are interpolated and smoothed to generate an airflow angle sequence. Spatial parameter mapping is performed between the airflow angle sequence and the pollutant heat map to generate evacuation wind direction data.

5. The tunnel ventilation control method based on environmental prediction as described in claim 4, characterized in that: The steps involve extracting the speed combinations and deflection angles of the tunnel ventilation fans from the evacuation wind direction data, filtering the speed combinations and deflection angles using a multi-objective reward mechanism, and outputting a set of fan control commands. Extract the standard wind direction angle from the evacuation wind direction data and record the grid coordinate range corresponding to the standard wind direction angle; Map the standard wind direction angles of each grid coordinate range to wind turbine parameters to obtain the speed combination and deflection angle; A multi-objective reward mechanism is used to score the rotation speed combination and deflection angle, and the rotation speed combination and deflection angle with the highest reward score are selected as candidate solutions for tunnel ventilation fans. The candidate solutions are converted into fan control commands and assigned to the corresponding tunnel ventilation fans according to the grid coordinate range, and the fan control command set is output.

6. The tunnel ventilation control method based on environmental prediction as described in claim 5, characterized in that: The reward scoring of speed combination and deflection angle through multi-objective reward mechanism refers to assigning reward weights to each speed combination and deflection angle based on the energy consumption, evacuation efficiency and safety objectives of the tunnel ventilation fan, and calculating the sum of the reward weights of each speed combination and deflection angle as the reward score.

7. The tunnel ventilation control method based on environmental prediction as described in claim 6, characterized in that: The steps involve adjusting the speed and deflection angle of the tunnel ventilation fan according to the fan control command set, guiding directional airflow, and monitoring the pollutant concentration in each tunnel ventilation area in real time using a smoke concentration sensor. The fan control command set is converted into PLC communication protocol signals and sent to the fan drive center of the corresponding tunnel ventilation fan; The fan drive center adjusts the fan speed and deflection angle of the tunnel ventilation fan to create directional airflow guidance; After directional airflow guidance, the pollutant concentration values ​​of each tunnel ventilation area are obtained by smoke concentration sensors at fixed time intervals.

8. The tunnel ventilation control method based on environmental prediction as described in claim 7, characterized in that: The step of comparing the predicted concentration values ​​in the pollutant heat map with the actual pollutant concentration values ​​to verify the effectiveness of tunnel ventilation control is as follows: Extract the predicted concentration value of each grid in the pollutant heatmap and record the grid coordinate range corresponding to the predicted concentration value; The predicted concentration values ​​and pollutant concentration values ​​are spatiotemporally aligned according to the grid coordinate range to generate pollutant data pairs. The predicted concentration value and the actual concentration value of pollutants in the data pair are compared point by point, the pollutant concentration difference is output, and a concentration threshold is set. If the pollutant concentration difference is positive and the absolute value of the pollutant concentration difference does not exceed the concentration threshold, then the ventilation control effect is deemed to be up to standard. If the pollutant concentration difference is negative, the grid coordinate range corresponding to the pollutant concentration value is marked as an abnormal area, and an early warning iteration mechanism is triggered.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the tunnel ventilation control method based on environmental prediction as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the tunnel ventilation control method based on environmental prediction as described in any one of claims 1 to 8.

Images