Tunnel variable frequency fan cooperative starting control method based on pre-scan air model

Abstract

This invention relates to the field of tunnel ventilation safety control technology, and discloses a method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging model. The method includes: generating a steady-state sequence data stream based on environmental state data and traffic flow video images; generating a warning trigger signal based on the steady-state sequence data stream; determining vehicle motion characteristic parameters and identifying multiple wind speed components; generating an effective driving airflow based on the multiple wind speed components and determining the diffusion mileage range of pollutants; analyzing the steady-state concentration sequence to generate the concentration rise rate and concentration change acceleration; adjusting the pre-scavenging airflow multiple coefficient to obtain the adjusted coefficient; generating a target airflow command; generating a critical anti-backflow wind speed based on tunnel hydraulic parameters and temperature gradient data; generating a limited-amplitude high-slope start-up command; and controlling the fans to execute control actions. This invention eliminates multi-source airflow interference, overcomes response lag defects, and achieves feedforward airflow compensation and blocking of high-temperature flue gas backflow in the early stages of pollution outbreaks.

Description

Technical Field

[0001] This invention relates to the field of tunnel ventilation safety control technology, specifically a method for coordinated start-up control of tunnel variable frequency fans based on a pre-sweeping air model. Background Technology

[0002] As a core facility for ensuring air quality inside tunnels, tunnel ventilation systems undertake multiple tasks, including removing vehicle exhaust fumes, diluting the concentration of harmful gases, maintaining visibility levels, and ventilating and dissipating heat during fires. In modern tunnel operation systems, the intelligence level and response efficiency of ventilation control systems have become important indicators for measuring the safe operation capability of tunnels.

[0003] Tunnel variable frequency fan coordinated start-up control technology is one of the core technologies of tunnel ventilation system. However, the existing technology has exposed many technical defects in practical applications, which seriously restricts the performance and reliability of tunnel ventilation system.

[0004] First, existing technology adopts a passive response mode, which only triggers the fan to start when the monitored data exceeds the warning threshold. There is a significant time lag between the generation of the alarm signal and the fan reaching the target speed, usually 15 to 30 seconds. During this period, the smoke has spread to a wider area, missing the best time for intervention and resulting in a significant reduction in ventilation effect.

[0005] Secondly, existing technologies ignore the diffusion path and direction characteristics of smoke in tunnels. Regardless of whether the fan is located upwind of the smoke diffusion direction, a uniform start-up logic is used, which cannot form an effective air curtain interception. The ventilation airflow and the smoke diffusion direction are difficult to form a counteracting relationship, reducing the smoke exhaust efficiency.

[0006] Furthermore, the existing air volume calculation methods are too crude, calculating the required air volume based only on the concentration data of a single monitoring point, without considering that the coverage area of ​​smoke diffusion is usually larger than the monitoring radius of a single sensor. The calculated air volume demand value is significantly lower than the actual required air volume, resulting in insufficient ventilation capacity.

[0007] Furthermore, when multiple monitoring points in the tunnel alarm simultaneously, existing technologies typically process each alarm point independently, without considering the cumulative effect of smoke diffusion paths in different areas. This makes it difficult to form a coherent and efficient smoke exhaust channel, resulting in a severe deficiency in multi-area collaborative control capabilities. Summary of the Invention

[0008] The purpose of this invention is to provide a method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model, so as to solve the problems mentioned in the background art.

[0009] To address the above problems, the present invention provides the following technical solution:

[0010] This invention provides a method for coordinated start-up control of variable frequency fans in tunnels based on a pre-scavenging air model. The method includes: acquiring environmental state data and traffic flow video images; generating a steady-state sequence data stream containing a steady-state concentration sequence based on the environmental state data; generating a warning trigger signal based on the steady-state sequence data stream; responding to the warning trigger signal, determining vehicle motion characteristic parameters based on the traffic flow video images; determining multiple wind speed components in the tunnel based on the environmental state data and the vehicle motion characteristic parameters; generating an effective driving airflow based on the multiple wind speed components; determining the diffusion mileage range of pollutants based on the effective driving airflow; and determining the interception wind within the diffusion mileage range. The system includes an interceptor fan and an auxiliary fan; it analyzes the steady-state concentration sequence to generate the concentration rise rate and concentration change acceleration, dynamically adjusts the pre-scavenging air volume ratio coefficient based on the concentration rise rate and concentration change acceleration, obtains the adjusted pre-scavenging air volume ratio coefficient, and generates a target air volume command based on the adjusted pre-scavenging air volume ratio coefficient; it acquires tunnel hydraulic parameters and temperature gradient data, generates a critical anti-backflow wind speed based on the tunnel hydraulic parameters and temperature gradient data, generates a limited-amplitude high-slope start command based on the critical anti-backflow wind speed, and controls the interceptor fan and the auxiliary fan to perform corresponding control actions based on the limited-amplitude high-slope start command.

[0011] As a further improvement of the present invention, the environmental state data includes carbon monoxide concentration data and visibility data. The step of generating a steady-state sequence data stream containing a steady-state concentration sequence based on the environmental state data, and generating an early warning trigger signal based on the steady-state sequence data stream, includes: performing dynamic noise reduction processing on the carbon monoxide concentration data and the visibility data based on a Kalman filter algorithm to establish a discrete state-space model; introducing a pre-defined matrix non-singularity test logic before the covariance matrix inversion step in the discrete state-space model to generate an optimized model; processing the environmental state data based on the optimized model to output a steady-state concentration sequence and a steady-state visibility sequence; generating a first monitoring signal based on the steady-state concentration sequence and a carbon monoxide concentration upper limit threshold, and generating a second monitoring signal based on the steady-state visibility sequence and a visibility lower limit threshold; and generating an early warning trigger signal based on the first monitoring signal or the second monitoring signal.

[0012] As a further improvement of the present invention, the vehicle motion characteristic parameters include vehicle density, average vehicle frontal area, and average vehicle speed. The step of determining the multiple wind speed components of the tunnel based on the environmental state data and the vehicle motion characteristic parameters includes: determining the natural base wind speed component and wind direction based on the environmental state data; generating a reference flow direction code based on the natural base wind speed component and the wind direction; determining the traffic piston equivalent pressure difference based on the vehicle density, the average vehicle frontal area, the average vehicle speed, and the tunnel cross-sectional area; generating a traffic piston wind component based on the traffic piston equivalent pressure difference; determining the real-time temperature at the current moment of the warning point corresponding to the warning trigger signal; determining the temperature difference between the real-time temperature and the absolute temperature of the normal environment; obtaining the local longitudinal slope angle of the section corresponding to the warning trigger signal; and generating a thermo-pressure buoyancy wind speed component based on the gravitational acceleration constant, the reference length of the reference section, the local longitudinal slope angle, and the temperature difference.

[0013] As a further improvement of the present invention, the step of generating an effective driving airflow based on the multi-wind speed components, determining the diffusion mileage range of pollutants based on the effective driving airflow, and determining the intercepting fan and auxiliary fan within the diffusion mileage range includes: generating an effective driving airflow based on the multi-wind speed components and the reference flow direction code; constructing a diffusion calculation model, determining the diffusion coverage radius based on the effective driving airflow and the diffusion calculation model; determining the diffusion mileage range based on the diffusion coverage radius; obtaining the physical location of each fan in space, and determining the intercepting fan and auxiliary fan based on the physical location in space and the diffusion mileage range.

[0014] As a further improvement of the present invention, the step of analyzing the steady-state concentration sequence to generate the concentration rise rate and concentration change acceleration, and dynamically adjusting the pre-scavenging air volume ratio coefficient based on the concentration rise rate and the concentration change acceleration to obtain the adjusted pre-scavenging air volume ratio coefficient includes: performing first-order derivative processing on the steady-state concentration sequence to obtain the concentration rise rate; performing second-order difference processing on adjacent concentration rise rates to obtain the concentration change acceleration; obtaining the tunnel cross-sectional area, determining the basic mapping value of the coefficient based on the concentration rise rate and the tunnel cross-sectional area; and dynamically adjusting the pre-scavenging air volume ratio coefficient based on the basic mapping value of the coefficient, the concentration rise rate, the concentration change acceleration, the change rate compensation weight coefficient, and the change acceleration compensation weight coefficient to obtain the adjusted pre-scavenging air volume ratio coefficient.

[0015] As a further improvement of the present invention, the step of generating a critical anti-backflow wind speed based on the tunnel hydraulic parameters and the temperature gradient data, and generating a limited-amplitude high-slope start-up command based on the critical anti-backflow wind speed, includes: determining the smoke front region based on the early warning trigger signal; obtaining the longitudinal absolute temperature difference of the smoke front region, and determining the hydraulic equivalent diameter based on the tunnel cross-sectional area; determining the critical anti-backflow wind speed based on the anti-backflow safety margin coefficient, the gravitational acceleration constant, the hydraulic equivalent diameter, the longitudinal absolute temperature difference, and the normal ambient absolute temperature; determining the steady-state operating wind speed based on the pre-sweeping air volume multiple coefficient, and generating a steady-state operating control command based on the steady-state operating wind speed and the critical anti-backflow wind speed; obtaining a limited-amplitude anti-backflow start-up curve, and generating a limited-amplitude high-slope start-up command based on the steady-state operating control command and the limited-amplitude anti-backflow start-up curve.

[0016] As a further improvement of the present invention, after generating the warning trigger signal, the method further includes: determining whether multiple warning trigger signals are generated in the tunnel; if so, obtaining the global environmental basic wind direction, determining the associated wind turbines corresponding to the warning trigger signals, performing spatial topology sorting on the associated wind turbines based on the global environmental basic wind direction, and obtaining wind turbine sorting information, the wind turbine sorting information including upstream and downstream groups; obtaining the initial smoke exhaust wind speed of the upstream group, and determining the average peak propagation wind speed based on the initial smoke exhaust wind speed; determining the straight-line physical spatial distance between the upstream group and the downstream group, and determining the physical propagation delay time based on the average peak propagation wind speed and the straight-line physical spatial distance; obtaining the initial dynamic pressure amplitude output by the upstream group, and determining the remaining effective wind pressure based on the initial dynamic pressure amplitude and the straight-line physical spatial distance.

[0017] As a further improvement of the present invention, the method further includes: obtaining the duration of the wind turbine startup; when the duration reaches the physical propagation delay time, controlling the variable frequency drive execution module corresponding to the downstream train to switch to constant output torque control mode; obtaining data on the increase in hysteretic wind pressure, and determining the frequency reduction amplitude required to absorb the excess increase in wind pressure based on the data on the increase in hysteretic wind pressure; and controlling the variable frequency drive execution module corresponding to the downstream train to perform a fine-tuning downward movement of the operating frequency based on the frequency reduction amplitude.

[0018] As a further improvement of the present invention, the method further includes: after controlling the intercepting fan and the auxiliary fan to perform corresponding control actions, acquiring the carbon monoxide concentration value and visibility value of the alarm monitoring point corresponding to the early warning trigger signal; generating a ventilation effect evaluation result based on the carbon monoxide concentration value and visibility value; if the ventilation effect evaluation result meets the preset reset condition, controlling the fan's operating frequency to return to the steady-state operating frequency; if the ventilation effect evaluation result does not meet the preset reset condition, generating a full-load operation command; and controlling the corresponding variable frequency drive execution module to perform the corresponding full-load operation based on the full-load operation command.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. This invention decouples the mixed wind speed in a confined space into a natural base wind speed component, a traffic piston wind component, and a thermo-pressure buoyancy wind speed component, eliminating the interference of traffic disturbances and thermo-buoyancy on a single wind measurement reading. This enables accurate identification of the pollution diffusion boundary and establishment of the target fan. Simultaneously, by performing differential calculations on the steady-state concentration sequence to obtain the concentration rise rate and change acceleration, the pre-scavenging air volume multiple coefficient is dynamically adjusted accordingly. This overcomes the response lag defect of traditional static threshold feedback control and achieves active feedforward air volume compensation in the early stages of a pollution outbreak.

[0021] 2. This invention constructs a critical anti-backflow control mechanism for high-temperature smoke backflow conditions during fires. Based on tunnel hydraulic parameters and longitudinal absolute temperature difference, the critical anti-backflow wind speed required to suppress the thermal buoyancy backflow layer is calculated, and a limited-amplitude high-slope start command is issued to the variable frequency drive execution module. This mechanism controls the intercepting fan to operate at a frequency exceeding the rated upper limit for a short period of time without triggering hardware overload protection, thereby establishing a high dynamic pressure air barrier within a short time window, effectively curbing the reverse spread of high-temperature smoke to the upstream of the tunnel.

[0022] 3. In handling concurrent smoke exhaust in multiple sections, this invention introduces aerodynamic impedance matching logic. By calculating the physical propagation delay time of the upstream air pressure peak to the downstream and the increase in lag pressure, the downstream variable frequency drive module is controlled to switch to constant output torque control at the airflow convergence node and the operating frequency is finely adjusted downward. This control method actively absorbs the excess air pressure at the intake end through flexible buffering at the electrical level, resolves the aerodynamic load impact when the upstream and downstream airflows converge, avoids smoke exhaust flow field turbulence and fan overcurrent tripping, and ensures the stability of continuous smoke exhaust throughout the entire area. Attached Figure Description

[0023] Figure 1 A system architecture diagram provided for embodiments of the present invention;

[0024] Figure 2This is a flowchart illustrating the specific implementation of the tunnel variable frequency fan coordinated start-up control method based on a pre-scavenging air model provided in an embodiment of the present invention.

[0025] Figure 3 The control effect comparison and verification diagram of the present invention is shown in (1), which is a comparison diagram of the environmental concentration decay trend during the smoke exhaust period, and (2) is a comparison diagram of the evolution of the bottom drive command of the variable frequency fan array. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] See attached document Figure 1 The present invention provides a method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model. During implementation, the corresponding control process is achieved through the cooperation of a multi-source state perception module, a data preprocessing module, an early warning judgment module, a diffusion model calculation module, a feedforward control module, an impedance matching module, and a variable frequency drive execution module.

[0028] The multi-source state perception module is deployed along the tunnel and at the tunnel entrance, including an environmental sensor array for acquiring carbon monoxide concentration data and visibility data, a wind speed sensor for acquiring basic environmental wind speed data, and a visual monitoring device for acquiring vehicle traffic images.

[0029] The data preprocessing module is communicatively connected to the multi-source state sensing module. The data preprocessing module receives raw time-series data collected by the environmental sensor array and wind speed sensor, and uses a filtering algorithm to remove high-frequency random disturbances and extract steady-state signal data.

[0030] The early warning judgment module is configured with a dynamic safety threshold. The early warning judgment module receives the steady-state sequence data stream after it has been processed by the data preprocessing module. When the steady-state sequence data stream exceeds the preset dynamic safety threshold, an early warning trigger signal is generated.

[0031] The diffusion model calculation module stores the cross-sectional area sequence matrix and the local longitudinal slope angle matrix along the tunnel. After receiving the early warning trigger signal, the module combines the vehicle density and speed information extracted by the visual monitoring equipment to calculate the natural base wind speed, traffic piston wind component and thermo-pressure buoyancy wind speed component, and then determines the estimated diffusion coverage radius of the smoke to delineate the polluted section.

[0032] The feedforward control module is connected to the diffusion model calculation module. The feedforward control module calculates the pre-sweeping air volume ratio coefficient based on the first and second derivatives of the steady-state signal data, and calculates the critical anti-backflow wind speed in combination with fluid dynamic parameters to generate the first drive control command.

[0033] The impedance matching module is designed for multi-region early warning conditions. It calculates the physical propagation delay time of the upstream airflow to the section where the downstream wind turbine is located, as well as the increase in lag pressure, and generates adaptive frequency reduction fine-tuning commands.

[0034] The variable frequency drive execution module is communicatively connected to the feedforward control module and the impedance matching module. Based on received commands, the variable frequency drive execution module adjusts the operating frequency of the tunnel ventilation fan array.

[0035] See attached document Figure 2 This invention provides a method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model, comprising the following steps:

[0036] S10, acquire environmental state data and traffic flow video images, generate a steady-state sequence data stream containing a steady-state concentration sequence based on the environmental state data, and generate an early warning trigger signal based on the steady-state sequence data stream;

[0037] Specifically, the multi-source state perception module collects environmental state data and traffic flow video images in the tunnel in real time. The data preprocessing module performs low-pass filtering on the raw time series data output by the environmental sensors to obtain a steady-state sequence data stream after eliminating transient disturbances. The early warning judgment module continuously compares the steady-state sequence data stream with the preset dynamic safety threshold. When the steady-state sequence data stream exceeds the preset dynamic safety threshold, an early warning trigger signal is generated and the collaborative control process is triggered.

[0038] S20, in response to the warning trigger signal, determine vehicle motion characteristic parameters based on the traffic flow video image, and determine multiple wind speed components of the tunnel based on the environmental state data and the vehicle motion characteristic parameters;

[0039] Specifically, after receiving the warning trigger signal, the diffusion model calculation module extracts the basic wind speed data from the environmental state data, converts the traffic flow video image into vehicle motion characteristic parameters, and calculates the natural basic wind speed component, traffic piston wind component, and thermal pressure buoyancy wind speed component by combining the preset tunnel spatial structure parameters.

[0040] S30, based on the multi-speed components, generate an effective driving airflow, determine the diffusion mileage range of pollutants based on the effective driving airflow, and determine the intercepting fan and auxiliary fan within the diffusion mileage range;

[0041] Specifically, the diffusion model calculation module superimposes each wind speed component to calculate the effective driving airflow, thereby delineating the diffusion mileage range of pollutants and establishing the intercepting and auxiliary fans within the target section;

[0042] S40, the steady-state concentration sequence is analyzed to generate the concentration rise rate and concentration change acceleration. The pre-scavenging air volume ratio coefficient is dynamically adjusted based on the concentration rise rate and concentration change acceleration to obtain the adjusted pre-scavenging air volume ratio coefficient. The target air volume command is generated based on the adjusted pre-scavenging air volume ratio coefficient.

[0043] Specifically, the feedforward control module performs differential calculations on the steady-state concentration sequence in the core pollution area to obtain the concentration rise rate and its acceleration. Based on the acceleration value, it dynamically adjusts the pre-scavenging air volume ratio coefficient and generates a target air volume command based on the adjusted pre-scavenging air volume ratio coefficient to execute preliminary fan control operations.

[0044] S50, acquire tunnel hydraulic parameters and temperature gradient data, generate critical anti-backflow wind speed based on the tunnel hydraulic parameters and temperature gradient data, and generate a limited high-slope start command based on the critical anti-backflow wind speed;

[0045] Specifically, the feedforward control module calculates the critical anti-backflow wind speed to prevent airflow from flowing in the opposite direction based on the tunnel hydraulic parameters and temperature gradient, and sends a limited high-slope start command to the variable frequency drive execution module. The variable frequency drive execution module of the intercepting fan increases the operating frequency to the set upper limit with the allowable preset acceleration within the safe torque range, maintains high dynamic pressure for a short time to establish a blocking air curtain, and then falls back to the steady-state operating frequency determined by the pre-sweeping air volume multiple coefficient.

[0046] S60, based on the limited high slope start command, control the intercepting fan and the auxiliary fan to perform corresponding control actions.

[0047] In this embodiment of the invention, when performing step S10, the following sub-steps are used to achieve the acquisition of multi-source heterogeneous data, time alignment, anti-interference processing, and anti-shake warning triggering:

[0048] First, a multi-dimensional heterogeneous data set within the tunnel and at entrance / exit nodes is synchronously collected and timestamped using a multi-source state perception module. Specifically, the multi-source state perception module includes an environmental sensor array, a wind speed sensor, and visual monitoring equipment. The environmental sensor array consists of carbon monoxide concentration sensors and visibility meters equidistantly distributed along the tunnel's longitudinal direction, while the wind speed sensor can be an ultrasonic anemometer.

[0049] During data acquisition, the environmental sensor array collects raw values ​​of carbon monoxide concentration and visibility at a preset sampling period. Preferably, the standard sampling period is set to 30 seconds; when an upward trend in concentration data is detected at any monitoring point, the system automatically shortens the sampling interval to a shorter period (e.g., 5 seconds) to improve the real-time performance of data acquisition under abnormal conditions. Simultaneously, wind speed sensors collect raw values ​​of ambient wind speed and natural wind direction, and visual monitoring equipment collects video stream data of vehicle traffic at tunnel entrances and along the roadways.

[0050] To address the issue of varying sampling periods across different hardware devices, the multi-source state perception module is equipped with a high-precision global clock. Upon receiving data packets from various underlying hardware sources, a synchronization timestamp corresponding to the current absolute system time is appended to the packet header. If network fluctuations cause short-term delays or data loss in some sensors, data from the previous valid period is reused via a zero-order hold logic. Ultimately, the system reassembles the discrete data into a multi-dimensional data frame sequence aligned along the time axis, providing a time reference for subsequent model calculations.

[0051] Then, the data preprocessing module performs low-pass filtering on the raw sequence data with timestamps to extract the steady-state sequence data stream that reflects the real trend of environmental changes.

[0052] Because high-speed vehicle traffic in tunnels generates piston-like turbulence, and heavy vehicles are prone to transient and concentrated exhaust emissions, the raw time-series data collected by the underlying sensors often contains high-frequency random noise. To avoid this noise causing oscillations in the control commands of the variable frequency drive execution module, the data preprocessing module, upon receiving the multi-dimensional data frame sequence, applies the Kalman filter algorithm for dynamic noise reduction processing for the carbon monoxide concentration data channel and the visibility data channel, respectively.

[0053] In its implementation, the data preprocessing module establishes a discrete state-space model with the previous state estimate and the current observation as input variables. It iteratively calculates the Kalman gain and continuously corrects the optimal state estimate for the current moment. To prevent covariance matrix inversion from failing under extreme conditions, a matrix non-singularity check is introduced before the inversion step: if the data matrix is ​​found to be trending towards singularity, a temporary switching to a moving average filtering mode is performed for degradation. After the above filtering process, the system removes high-frequency spike noise and outputs a smooth and continuous steady-state concentration and visibility sequence. For the basic state equation, observation equation, and conventional parameter initialization configuration of the covariance matrix in the Kalman filter algorithm, those skilled in the art can perform routine calculations and calibrations based on the hardware manual of the specific tunnel sensor. The basic mathematical derivation and code implementation are well-known technologies in this field and will not be elaborated upon here.

[0054] At this point, the early warning judgment module performs a dual-condition anti-shake judgment based on the steady-state sequence data stream processed by the data preprocessing module and generates an early warning trigger signal. Specifically, the early warning judgment module is internally configured with dynamic safety thresholds including an upper limit threshold for carbon monoxide concentration and a lower limit threshold for visibility. In the case of a tunnel application, the upper limit threshold for carbon monoxide concentration can be set to 50 ppm, and the lower limit threshold for visibility can be set to 500 m. In this embodiment of the invention, the system can dynamically switch the safety thresholds according to time period and traffic flow. For example, the upper limit threshold for carbon monoxide concentration can be appropriately lowered during peak hours with heavy traffic, and then reverted to the baseline threshold during off-peak hours at night.

[0055] The early warning judgment module continuously receives steady-state sequence data streams using a sliding time window. These streams include steady-state concentration and visibility sequences, and are compared frame-by-frame with the corresponding dynamic safety thresholds. To prevent data fluctuations at threshold edges, a time hysteresis anti-jitter mechanism is introduced: when the system is in short-cycle sampling mode, the time window is preferably set to 10 seconds; when in standard sampling cycle mode, the time window is preferably set to at least two consecutive sampling cycles. A first monitoring signal is generated when the steady-state concentration sequence value is consistently greater than the upper limit threshold for carbon monoxide concentration within the time window; or a second monitoring signal is generated when the steady-state visibility sequence value is consistently less than the lower limit threshold for visibility within the time window. When either of these signals is detected, the steady-state sequence data stream is determined to exceed the preset dynamic safety threshold, and an early warning trigger signal is generated.

[0056] Furthermore, to improve the accuracy of early warnings and avoid false triggering, once any steady-state sequence data reaches the early warning condition, the early warning judgment module performs auxiliary verification by combining the changing trends of similar parameters at adjacent monitoring points, or the synchronous changing trends of another type of environmental parameter. When the steady-state sequence data stream of the target monitoring point continuously exceeds the preset dynamic safety threshold and adjacent monitoring points or another type of environmental parameter show the same abnormal trend, the early warning trigger signal is confirmed to be valid. Subsequently, the physical coordinate mileage of the sensor that exceeded the limit is extracted, and an early warning trigger signal containing the alarm time, alarm coordinate mileage, and alarm source data type is generated and sent to the diffusion model calculation module.

[0057] In step S20, the mixed wind speed, which is affected by multiple physical factors, is refined and separated into three independent physical vectors to provide data support for the subsequent delineation of pollution boundaries:

[0058] First, after receiving the warning trigger signal sent by the warning judgment module, the diffusion model calculation module extracts the environmental basic wind speed data from the steady-state sequence data stream and calculates the natural basic wind speed component.

[0059] To eliminate the interference of local traffic disturbances and thermal buoyancy on instantaneous wind speed readings, the diffusion model calculation module performs time-domain analysis on longitudinal wind speed data after receiving the warning trigger signal, and calculates the average wind speed value within the time window (preferably 10s) before the trigger as the natural base wind speed component.

[0060] Simultaneously, the module records the positive and negative algebraic signs of the natural basic wind speed components. Taking the direction from the tunnel entrance to the exit as the positive reference direction, a reference flow direction code is established: a positive wind speed value is recorded as positive airflow, coded as 1; a negative wind speed value is recorded as reverse airflow, coded as -1; and a wind speed value of 0 indicates that the airflow is stationary, coded as 0.

[0061] At this point, the diffusion model calculation module converts the traffic flow video images acquired by the visual monitoring equipment into vehicle motion feature parameters, and combines them with an aerodynamic model to deduce the traffic piston wind component. Specifically, the diffusion model calculation module calls a visual target detection and tracking algorithm to analyze the traffic flow video frames frame by frame, extracts the real-time vehicle density and average vehicle speed of the current warning section, and queries to obtain the tunnel cross-sectional area of ​​the corresponding section.

[0062] Simultaneously, based on the principle of equivalent transformation of fluid dynamics resistance, the traffic piston wind component is calculated according to vehicle density, average vehicle speed, and tunnel cross-sectional area. Specifically, the equivalent pressure difference of the traffic piston is first calculated based on vehicle density, average vehicle frontal area, average vehicle speed, and tunnel cross-sectional area. The number of vehicles in the reference section is determined according to vehicle density and reference section length. The blockage ratio is determined based on the number of vehicles, average vehicle frontal area, and tunnel cross-sectional area. The equivalent pressure difference of the traffic piston is then calculated by combining the square of the average vehicle speed. Subsequently, based on the conversion relationship between this pressure difference and air density, the traffic piston wind component is obtained.

[0063] When the vehicle's average frontal area or air density is defaulted, preset calibration values ​​are used in the calculation. To avoid calculation overflow caused by the tunnel cross-sectional area abnormally approaching 0 during division operations, the module has a built-in safety baseline cross-sectional area constant (e.g., the minimum effective cross-sectional area of ​​a conventional tunnel is 30m²). 2 ( ); When the cross-sectional area data is abnormal, a fault-tolerant mechanism is implemented, replacing the denominator of the formula with this constant. For the aforementioned visual target detection feature extraction algorithm and the mathematical derivation of the basic aerodynamic drag formula, those skilled in the art can refer to mature computer vision frameworks and tunnel ventilation design specifications for conventional configuration; these are well-known technologies in the field and will not be elaborated upon here.

[0064] At this point, the diffusion model calculation module extracts the temperature data of the warning point and calculates the thermal pressure buoyancy wind speed component caused by the temperature difference slope in conjunction with the local longitudinal slope angle matrix. Specifically, the diffusion model calculation module extracts the real-time temperature of the warning point at the current moment, calculates the temperature difference between it and the absolute temperature of the normal environment, and retrieves the local longitudinal slope angle of the corresponding section in conjunction with the alarm coordinate mileage.

[0065] Based on the principle of energy conservation and the law of thermal buoyancy, the gravitational acceleration constant, the reference length of the benchmark section, the sine value of the local longitudinal slope angle, and the ratio of the absolute temperature difference to the absolute temperature of the normal environment are calculated. The square root of the product of the above four parameters is obtained and multiplied by the thermal buoyancy resistance compensation coefficient to obtain the thermal pressure buoyancy wind speed component.

[0066] The thermal buoyancy resistance compensation coefficient ranges from 0.65 to 0.85, and is determined based on the roughness of the tunnel wall around the warning point and the friction constant along the tunnel. The reference length of the benchmark section is set based on the typical spread characteristics of fire smoke plumes in the tunnel, and ranges from 100m to 200m. The typical benchmark value for the absolute temperature of the normal environment is 293.15K (if the reading abnormally approaches absolute 0°, this benchmark value is forcibly adopted to ensure operational stability).

[0067] After obtaining the value of the thermal pressure buoyancy wind speed component, the direction of the component is determined according to the positive and negative directions of the local longitudinal slope angle and the relationship between the temperature at the warning point and the normal ambient temperature. When the local temperature difference is positive and the slope rise direction corresponding to the local longitudinal slope angle is consistent with the tunnel's positive direction, it is recorded as positive; when the local temperature difference is positive and the slope rise direction is opposite to the tunnel's positive direction, it is recorded as negative; when the local temperature difference is close to 0 or lower than the preset lower limit of temperature difference, the component is treated as 0 or the bottom line compensation value.

[0068] During step S30, the multiphysics airflow parameters are converted into quantified pollution boundaries, and the corresponding operating modes of the physical execution hardware are determined accordingly.

[0069] First, the diffusion model calculation module performs vector superposition of the decoupled multi-wind speed components to calculate the effective driving airflow that actually drives the diffusion of pollutants.

[0070] The diffusion model calculation module determines the physical orientation of the traffic piston wind component and the thermo-buoyancy wind speed component based on the baseline flow direction code and assigns them corresponding positive and negative algebraic signs. Subsequently, the natural base wind speed component, the traffic piston wind component, and the thermo-buoyancy wind speed component are algebraically added to obtain the calculated value of the effective driving airflow with directional signs.

[0071] The system uses the positive or negative sign of the calculated value to determine the downstream diffusion direction of pollutants. Simultaneously, it takes the absolute value of the calculated value and uses the resulting velocity amplitude as the effective driving airflow in subsequent diffusion distance calculations. This preserves the basis for the diffusion direction and avoids negative velocity inputs in the distance calculation model.

[0072] Then, the diffusion model calculation module estimates the diffusion coverage radius based on the dynamic calculation of the effective driving airflow and delineates the start and end mileage coordinates of the polluted section.

[0073] The diffusion model computation module combines spatial geometric parameters and the time window dimension to construct a diffusion computation model, the specific formula of which is:

[0074] ;

[0075] In the formula, Indicates the estimated diffusion coverage radius. This represents the geometric diffusion constant used to compensate for the nonlinear diffusion expansion of irregular pipe walls, preferably... The value ranges from 1.2 to 1.5. For sections with good ventilation, a lower value closer to 1.2 is used, while for sections with poor ventilation or local obstructions, a higher value closer to 1.5 is used. Indicates effective driving airflow. This indicates the pre-scavenging time window, and the baseline value can be set to 60 seconds.

[0076] Based on the calculated estimated diffusion coverage radius, and using the alarm coordinate mileage as the spatial base point, the starting and ending mileage coordinate range of the polluted section is calculated by extending the corresponding physical length along the actual airflow propagation direction. Before outputting the coordinates, the system performs an out-of-bounds check: if the calculated starting or ending coordinates exceed the actual physical mileage of the tunnel entrance and exit, the actual entrance and exit mileage is automatically used as the cutoff boundary.

[0077] Finally, the diffusion model calculation module dynamically groups the target fans in the preset three-dimensional tunnel topology based on the starting and ending mileage coordinates of the polluted section to determine the intercepting fans and auxiliary fans.

[0078] The system retrieves the starting and ending mileage coordinates of the polluted section and the surrounding variable frequency wind turbine physical nodes, and executes a three-level grouping strategy: the first wind turbine located at the starting boundary of the upwind direction of the polluted area is marked as an interceptor wind turbine; adjacent wind turbines located inside the polluted section and downstream of the interceptor wind turbine are uniformly marked as auxiliary wind turbines; and wind turbines far away from the diffusion path and unaffected are marked as standby wind turbines.

[0079] After completing the role definition, issue differentiated response time constraint instructions: the interceptor fan has the highest response priority and must enter the start response state and reach the set acceleration slope within 5 seconds after receiving the instruction. The time for it to reach the target speed is executed according to the smooth acceleration time issued by the feedforward control module; the auxiliary fan must enter the start response state within 10 seconds; the standby fan must enter the standby or low-speed preparation state within 15 seconds.

[0080] In this embodiment of the invention, the static control mechanism based on lookup tables is further adjusted to a feedforward dynamic control with trend prediction capabilities to cope with the nonlinear diffusion of smoke caused by sudden pollution or deflagration:

[0081] First, the feedforward control module performs differential calculations on the steady-state concentration sequence of the pollution core area to obtain the concentration rise rate and its acceleration.

[0082] The feedforward control module performs differential operations on the steady-state carbon monoxide concentration sequence of the core pollution area, including the early warning trigger point. Specifically, it calculates the ratio of the concentration difference between the current sampling period and the previous sampling period to the time, and obtains the steady-state first derivative to characterize the rate of concentration increase. Based on this, it performs a second difference on adjacent first derivatives to obtain the steady-state second derivative to characterize the acceleration of concentration change.

[0083] Then, the feedforward control module dynamically calculates the pre-scavenging air volume ratio coefficient based on the concentration rise rate and its change acceleration.

[0084] The feedforward control module introduces two evaluation indicators: carbon monoxide concentration rise rate and tunnel cross-sectional area, to determine the basic mapping range of the pre-scavenging air volume ratio coefficient. When the concentration rise rate exceeds the first preset rise rate (e.g., 5 ppm / s) and the tunnel cross-sectional area exceeds the first preset area, the basic mapping value of the coefficient is taken as close to the upper limit of the basic mapping range of the pre-scavenging air volume ratio coefficient (e.g., 1.5). When the concentration rise rate is lower than the second preset rise rate (e.g., 2 ppm / s) and the tunnel cross-sectional area is lower than the second preset area, the basic mapping value of the coefficient is taken as close to the lower limit of the basic mapping range of the pre-scavenging air volume ratio coefficient (e.g., 1.2). The first and second preset areas are cross-sectional area thresholds pre-calibrated and stored based on the tunnel design cross-sectional area, ventilation design parameters, and historical operating data, with the first preset area being greater than the second preset area. The first preset area is used to determine the operating conditions of large-section tunnels, and the second preset area is used to determine the operating conditions of small-section tunnels.

[0085] After establishing the basic mapping range, the feedforward control module executes the following pre-scavenging coefficient calculation model:

[0086] ;

[0087] In the formula, This represents the pre-sweeping air volume ratio factor, the value of which is limited to the range of 1.2 to 1.5 by hardware safety logic; This represents the aforementioned basic mapping value; This represents the steady-state concentration value after smoothing. It is a time variable; The rate of increase in concentration; The acceleration due to concentration change; The preferred range for the rate-of-change compensation weighting coefficient is 0.1 to 0.3. The preferred range for the weighting coefficients to compensate for the changing acceleration is 0.05 to 0.15. (Maximum value function) This is used to set the acceleration compensation term to 0 when the concentration change shows a decreasing trend.

[0088] Before being substituted into the model, the feedforward control module normalizes the concentration rise rate and concentration change acceleration. Specifically, it divides the concentration rise rate by the first reference rise rate and the concentration change acceleration by the first reference acceleration. These reference values ​​are pre-calibrated based on historical tunnel pollution data, sensor sampling periods, and motor safety output boundaries. The corresponding weighting coefficients are used as dimensionless weights in the calculation to ensure that the weighted sum of all items does not exceed the maximum safe output torque limit of the motor.

[0089] Finally, the feedforward control module issues a target air volume command based on the adjusted pre-scavenging air volume multiplier, guiding the interceptor fan to enter the pre-scavenging operation state.

[0090] The feedforward control module calculates the standard air volume based on the absolute value of the smoke concentration at the current monitoring point and the tunnel cross-sectional area. Then, it multiplies the adjusted pre-scavenging air volume multiplier by the standard air volume to obtain the final target pre-scavenging air volume. When the concentration change acceleration is greater than 0, the second derivative compensation term in the model is activated, causing the target pre-scavenging air volume to be amplified by feedforward.

[0091] After receiving the target airflow command, the variable frequency drive of the interceptor fan calls its internal smoothing curve control algorithm to gradually increase the rotational speed from 0 to the target rotational speed corresponding to the target pre-sweeping airflow. Preferably, the smoothing speed increase time is set to 10 seconds to avoid sudden torque changes causing shear impact on the fan's mechanical main shaft structure.

[0092] During step S50, an aerodynamic barrier is further established to prevent the smoke from spreading upstream, and a specific startup sequence is executed within the physical safety boundary of the hardware device:

[0093] First, the feedforward control module combines tunnel hydraulic parameters and temperature gradient data, and uses the Froude number conservation principle to deduce the critical anti-backflow wind speed to prevent airflow from flowing in the opposite direction. Specifically, the feedforward control module extracts the core highest temperature of the warning point and the undisturbed temperature of the upwind section to calculate the longitudinal absolute temperature difference in the smoke front area; at the same time, it determines the hydraulic equivalent diameter of the current section based on the ratio of four times the tunnel cross-sectional area to the wetted perimeter.

[0094] Based on the dimensionless Froude number model balancing thermal buoyancy and inertial force, the critical anti-backflow wind speed is calculated using the following formula:

[0095]

[0096] In the formula, Indicates the critical wind speed for preventing backflow; This represents the safety margin factor for preventing backflow, with a preferred range of 1.05 to 1.15; It is the gravitational acceleration constant; The diameter is the hydraulic equivalent. The longitudinal absolute temperature difference in the smoke front region; The absolute temperature is the normal ambient temperature. In addition, the system also sets a critical wind speed baseline protection value (e.g., 1.5 m / s). When the calculated result is lower than this baseline protection value, the baseline protection value is forcibly output to prevent non-fire-related pollution from causing the temperature difference to approach 0 and triggering a dead zone in subsequent calculations.

[0097] Then, the feedforward control module compares the wind speed command parameters and sends the amplitude-limiting high-slope start command and pressure-holding parameters to the frequency converter drive execution module.

[0098] The feedforward control module converts the target pre-sweeping air volume into a steady-state operating wind speed based on the adjusted pre-sweeping air volume multiplier coefficient, and cross-checks it with the critical anti-backflow wind speed to generate the corresponding steady-state operating control command: when the steady-state operating wind speed is lower than the critical anti-backflow wind speed, the critical anti-backflow wind speed is forcibly used as the benchmark control target; otherwise, the steady-state operating wind speed continues to be used.

[0099] To quickly establish an air curtain, the feedforward control module sends a limited high-slope start command containing specific timing logic to the interceptor fan. The command includes the final steady-state target operating frequency, the over-rated upper limit frequency parameter, and the pressure holding time setting value.

[0100] Finally, the variable frequency drive execution module performs a transient forced frequency reduction action sequence within the safe torque boundary to complete the establishment and smooth transition of the blocking air curtain.

[0101] After receiving the high-slope start command, the variable frequency drive execution module at the interception position calls the internally configured limit-limit anti-backflow start curve. The limit-limit anti-backflow start curve is a frequency-time control curve pre-calibrated according to the rated parameters of the variable frequency fan, motor, and variable frequency drive. This curve includes a rapid frequency increase section, a short-time pressure holding section, and a smooth fallback section, which are used to quickly establish a blocking air curtain without triggering electrical protection.

[0102] Under the premise of real-time monitoring of the main circuit output current and motor stator torque, and determining that over-limit protection is not triggered, the operating frequency is increased to the upper limit of the rated frequency by the preset electrical acceleration limit allowed by the equipment. The preset electrical acceleration limit refers to the maximum frequency rise slope allowed by the frequency converter drive execution module under the condition that the main circuit overcurrent protection and motor torque over-limit protection are not triggered. It can be pre-calibrated according to the rated power of the motor, the allowable output current of the frequency converter, and the mechanical inertia of the fan. This upper limit frequency is set according to the short-time overload capacity of the motor, and is typically 110% of the rated frequency (usually 50Hz), i.e., 55Hz.

[0103] The fan continues to operate in the over-frequency band. When the cumulative operating time reaches the pressure holding time set by the command (preferably 60s to 90s), it is determined that the airflow barrier has been stably established. Subsequently, the variable frequency drive execution module switches to the smooth deceleration ramp mode, gradually reducing the operating frequency and stabilizing it at the steady-state target operating frequency.

[0104] In this embodiment of the invention, the experience-based fixed timing combination is also adjusted to an adaptive dynamic spatial propagation calculation to reduce aerodynamic disturbances caused by concurrent smoke exhaust at multiple points in long tunnels:

[0105] First, the system performs flow field topology mapping for multi-segment concurrent early warning conditions, dividing the variable frequency wind turbine group into upstream and downstream groups.

[0106] In practical applications, real-time monitoring is used to determine whether warning trigger signals have been generated in multiple sections of the tunnel. If multiple sections are under concurrent warning conditions, the impedance matching module receives the start and end mileage coordinates of each section and extracts the global environmental basic wind direction. Using this wind direction as the reference axis, the polluted area under warning and its associated wind turbines are spatially topologically sorted.

[0107] The wind turbines in the warning area located upwind of the overall airflow are designated as upstream groups; those located downwind are designated as downstream groups. A master-slave spatial relationship is established through a flow field topology mapping mechanism to avoid flow field disturbances caused by uncoordinated and unsynchronized startup of upstream and downstream turbines.

[0108] Then, the impedance matching module calculates the physical propagation delay time of the air pressure wave peak to the downstream train based on the upstream cross-section wind speed and the physical distance between the upstream and downstream spaces.

[0109] The impedance matching module extracts the initial exhaust velocity of the upstream train and performs resistance compensation and attenuation conversion based on the friction coefficient along the tunnel wall to obtain the average peak propagation velocity. At the same time, it extracts the coordinates of the exhaust end of the upstream train and the intake end of the downstream train to obtain the straight-line physical spatial distance between them.

[0110] The physical propagation delay time of the air pressure wave crest to the downstream train formation is calculated by dividing the straight-line physical spatial distance by the average wave crest propagation wind speed. Preferably, the baseline value of the physical propagation delay time is limited to 5 seconds, taking into account the typical arrangement spacing of tunnel ventilation fans.

[0111] To prevent the average peak propagation wind speed from approaching zero and causing division overflow, the module has a built-in safety mandatory coordination time threshold (e.g., 30 seconds). When the calculated delay time exceeds this threshold, a fault tolerance mechanism is forcibly activated, using the independent start-up determination of the downstream fan as the control basis.

[0112] Finally, based on the pipeline resistance attenuation characteristics, the impedance matching module quantifies the hysteresis pressure increase when the gas flow pressure peak reaches the downstream group.

[0113] The impedance matching module extracts the initial dynamic pressure amplitude of the upstream group output. Based on Bernoulli's equation and the pipeline friction resistance model, it calculates the remaining effective pressure of the upstream initial exhaust gas pressure after overcoming the frictional resistance of the aforementioned straight physical space distance.

[0114] The remaining effective wind pressure is the amount of lag pressure increase that the downstream fan inlet will soon experience. The system records and preloads this parameter, providing a clear benchmark for the subsequent adaptive electrical unloading and impedance flexible matching of the actuators.

[0115] In this embodiment of the invention, the load impact caused by airflow convergence is mitigated at the electrical level, and a closed loop is established for evaluating the ventilation effect, issuing warnings, resetting the status, and forcing smoke extraction under extreme conditions after a preset time period of pre-sweeping and coordinated smoke extraction operation.

[0116] Specifically, through automatic timing via an internal clock, when the accumulated internal clock reaches the physical propagation delay time node, the downstream variable frequency drive execution module automatically switches the underlying drive algorithm to constant output torque control mode. In this mode, the variable frequency drive execution module limits the electromagnetic torque output at the motor shaft end as the primary control target, allowing the stator operating frequency to fluctuate within a preset tolerance range. This mechanism enables the mechanical rotating body to flexibly adapt to the sudden increase in positive pressure at the intake end, avoiding triggering overcurrent trip protection.

[0117] Meanwhile, the variable frequency drive execution module extracts the data on the increase in lag wind pressure and, based on the similarity law that the wind pressure of the fan is positively correlated with the square of the impeller speed, calculates the frequency reduction amplitude required to absorb the excess wind pressure increase.

[0118] During the fusion window period of receiving the upstream airflow peak, the downstream fan performs a fine-tuning downward adjustment of its operating frequency. Preferably, the frequency reduction is limited to the range of 2Hz to 5Hz. This frequency reduction mechanism promotes smooth fusion of airflows at the interface, forming a continuous smoke exhaust channel.

[0119] After a preset time period (preferably 30 seconds) of pre-scavenging and coordinated smoke extraction, the early warning judgment module continuously reads the carbon monoxide concentration and visibility values ​​at the alarm monitoring points to assess the ventilation effect. The preset time period is the continuous monitoring period used by the system to assess the ventilation effect after pre-scavenging and coordinated smoke extraction. It can be preset according to the tunnel length, fan spacing, ventilation design parameters, and historical operating data, for example, preferably 30 seconds. If the concentration value drops below the concentration drop target value (typically 40 ppm) and the visibility value rises above the visibility rise target value (typically 600 m), the reset condition is met, and the alarm is deactivated.

[0120] After confirming that the alarm has been cleared, the variable frequency drive execution module calls the internal soft deceleration ramp curve to steadily reduce the motor operating frequency back to the normal maintenance speed state (typically 15Hz), and the system returns to the normal monitoring state.

[0121] If the warning and judgment module determines that the environmental values ​​continue to deteriorate, the system will automatically switch to the forced smoke extraction mode, which includes external hardware linkage.

[0122] If, during the aforementioned ventilation effect assessment, the carbon monoxide concentration continues to rise or the visibility continues to decrease, it is determined that the pre-sweeping and coordinated smoke extraction effects have not met the preset requirements. The system automatically switches to forced smoke extraction mode and issues a full-load operation command to the associated fans. The variable frequency drive module increases the fan speed to the rated continuous maximum operating frequency (typically corresponding to 50Hz), allowing the fans to operate continuously at rated power. The 55Hz frequency during the high-slope start-up phase of the interceptor fan is only used for short-term pressure maintenance to establish an air curtain and is not used as the continuous operating frequency for this mode.

[0123] Simultaneously, the system triggers an external hardware linkage mechanism: controlling the tunnel entrance indicator light to switch to a red, no-entry state; activating the internal broadcast system to continuously play evacuation prompts; and establishing a two-way communication link with on-site personnel's mobile terminals to issue evacuation instructions. The linkage mechanism continues to operate until the steady-state concentration sequence falls below the upper limit threshold of carbon monoxide concentration and the steady-state visibility sequence rises above the lower limit threshold of visibility, at which point a reset operation can be performed.

[0124] The technical solution of the present invention will be described in detail below with reference to specific application embodiments:

[0125] In an embodiment of the invention, a one-way three-lane highway tunnel is 3000m long and has a cross-sectional area of ​​80m². 2 hydraulic equivalent diameter =7.5m, with a local longitudinal slope of +2%, and a fan array arranged every 200m inside the tunnel, with a normal ambient absolute temperature =293.15K (20℃), the normal carbon monoxide concentration is 10ppm.

[0126] During a peak hour, a heavy-duty diesel truck experienced a mechanical failure at 1000m in the tunnel (mileage coordinates: K1000), resulting in continuous exhaust emissions and engine overheating.

[0127] S10, Status Awareness and Early Warning Triggering

[0128] Data Acquisition and Filtering: The environmental sensor array acquires data at K1000 with a 30s cycle. Due to the violent oscillation of exhaust gas turbulence between 45 and 75, the raw concentration data is processed by the data preprocessing module using Kalman filtering to extract smooth steady-state sequence data.

[0129] Warning Trigger: If the steady-state carbon monoxide concentration sequence exceeds the dynamic upper limit threshold of 50 ppm for two consecutive periods (within 60 seconds), reaching 65 ppm, the warning judgment module generates a warning trigger signal, and the alarm coordinate is set to K1000.

[0130] S20, Multi-source wind speed decoupling calculation

[0131] Natural baseline wind speed component: Extract data from 10 seconds prior to the warning to calculate the natural baseline wind speed component. =1.5m / s (positive direction).

[0132] Traffic piston wind component: Visual equipment measured a high vehicle density at this time, with an average vehicle speed of 40 km / h. Based on the cross-sectional area and aerodynamic model, the traffic piston wind component was calculated. =2.0m / s (positive direction).

[0133] Thermo-pressure buoyancy wind speed component: The highest core temperature measured at K1000 was 50℃ (323.15K), and the longitudinal absolute temperature difference in the smoke front region. =30K, longitudinal slope upward, calculate the thermal pressure buoyancy wind speed component. =0.5m / s (positive direction).

[0134] S30, Pollution Boundary Delineation and Role Assignment

[0135] Effective driving airflow: The above components are superimposed to form the effective driving airflow. =1.5+2.0+0.5=4.0m / s.

[0136] Diffusion radius calculation: Extracting the geometric diffusion constant =1.25, pre-scavenging time window =60s, call the diffusion calculation model formula Calculate the estimated diffusion coverage radius for:

[0137] ;

[0138] Polluted section: extending from K1000 to K1300.

[0139] Turbine grouping: The turbine at K950 is designated as the interceptor turbine (upstream boundary); the turbine at K1150 is designated as the auxiliary turbine; and the turbine at K1400 is designated as the downstream group (impedance matching node).

[0140] S40, Dynamic Calculation of Feedforward Coefficients

[0141] Differential operation: for the smoothed steady-state concentration value Perform calculations to obtain the current rate of increase in concentration. =6ppm / s, and the acceleration of concentration change =2ppm / s 2 .

[0142] Coefficient calculation: Extracting basic mapping values =1.3. Call the pre-scavenging coefficient calculation model formula. .

[0143] The above differential values ​​are normalized (the first reference rise rate is set to 5 ppm / s, and the first reference acceleration is 1 ppm / s²), and the rate of change compensation weighting coefficient is substituted into the values. =0.2, weighting coefficient for changing acceleration compensation =0.1, calculate the pre-scavenging air volume ratio factor. :

[0144] ;

[0145] This value triggers the hardware safety logic, and the system limits the pre-sweep air volume multiplier to the upper limit of 1.5 times.

[0146] S50, Calculation of Critical Anti-backflow Wind Speed ​​and Establishment of Air Curtain

[0147] Critical wind speed derivation: Extracting the safety margin coefficient for backflow prevention =1.1, gravitational acceleration constant =9.81m / s 2 Call the model formula based on the Froude number conservation principle. Calculate the critical backflow wind speed :

[0148] ;

[0149] Interception Execution: Upon receiving the limited high-slope start command, the K950 intercept fan will increase its operating frequency to 55Hz (110% overload) within 5 seconds with the set upper limit acceleration. It will maintain the set pressure holding time in this frequency band for 60 seconds to establish an air curtain with a wind speed greater than 3.01m / s, preventing the airflow from flowing upstream of the K1000. Then, it will smoothly return to the steady-state operating frequency (e.g., 45Hz) determined by the pre-sweeping air volume ratio factor.

[0150] S60, Physical Propagation Delay and Impedance Matching

[0151] Delay time calculation: The straight-line physical spatial distance between K1400 (downstream intake end) and K950 (upstream exhaust end) is extracted to be 450m. Combining the friction loss characteristics, the equivalent value of the average peak propagation wind speed is estimated to be the aforementioned effective driving airflow of 4.0m / s. The physical propagation delay time is then calculated.

[0152] Physical propagation delay time s;

[0153] Wind pressure compensation: Based on the pipeline friction model, the remaining effective wind pressure after the airflow overcomes the friction of the physical distance in the straight line is calculated, and the amount of lag wind pressure increase that the downstream train will bear is quantified.

[0154] At 112.5 seconds after the interceptor fan starts (i.e., reaching the physical propagation delay time node), the K1400 fan switches to constant output torque control state, and its operating frequency is slightly adjusted down by 3Hz according to the increase in lagging wind pressure. This mechanism effectively absorbs the positive pressure impact transmitted from upstream to this section, maintains the flow field stability of the converging smoke exhaust section, and avoids triggering overcurrent trip protection. After the preset time period of pre-sweeping and coordinated smoke exhaust operation, the system evaluates the ventilation effect and continuously reads the carbon monoxide concentration and visibility values ​​at the alarm monitoring points. When the steady-state sequence data stream falls back to the preset dynamic safety threshold range, the variable frequency drive execution module performs a smooth speed reduction reset operation, and the system returns to the 15Hz daily monitoring state.

[0155] Experimental verification and effect comparison:

[0156] To verify the effectiveness of the technical solution of this invention, a comparative experiment was conducted in the aforementioned real tunnel environment. The comparison was between the traditional proportional-integral-derivative (PID) feedback control method and the cooperative control method provided in the embodiments of this invention.

[0157] Evaluation indicators:

[0158] Peak carbon monoxide concentration: Used to assess the system's ability to contain environmental parameter deterioration. Pollution elimination time: The time required for the carbon monoxide concentration to drop below the safe threshold (40 ppm) from the triggering of the warning. Fan frequency response stability: Used to assess the actuator's flow field adaptability and electrical stability.

[0159] Comparative analysis of experimental data and visualization:

[0160] See attached document Figure 3 , Figure 3 The system control effect comparison and verification diagram reflects the carbon monoxide concentration decay trend during the smoke exhaust period after the warning is triggered, as well as the evolution of the underlying drive commands of the variable frequency fan array as the control execution time increases.

[0161] Regarding the pollution spread range, the pollution range under the traditional PID feedback control method extends to the K1000 to K1500 section, with a spread distance of up to 500m; the collaborative control method of the present invention controls the pollution section within the K1000 to K1300 range, limiting the spread distance to 300m, effectively reducing the affected area by 40%.

[0162] Regarding the peak concentration of carbon monoxide and the elimination time, combined with Figure 3 The comparison chart (1) of the environmental concentration decay trend during the smoke emission period shows the change trend of carbon monoxide concentration in the core pollution area of ​​the tunnel after the warning was triggered. The black dashed line with a circle in the figure represents the traditional feedback control method. Because it relies on feedback adjustment after the threshold is exceeded, and the upstream and downstream fans start simultaneously, causing mutual interference in the flow field, the system response is delayed. The peak carbon monoxide concentration rises to 88 ppm and it takes 315 seconds for the concentration to fall back to the safe range. The black solid line with a square in the figure represents the collaborative control method of this invention. Based on the concentration change acceleration feedforward mechanism, the peak carbon monoxide concentration is controlled at about 66 ppm, and the concentration decay rate is significantly accelerated. It falls back to below the safe warning threshold boundary after about 185 seconds.

[0163] In terms of system operational stability and frequency response, combined with Figure 3 The diagram (2) comparing the evolution of the underlying drive commands of the variable frequency fan array shows the operating frequency adjustment curve of the core area interceptor fan. The black dashed line marked with a circle represents the traditional feedback control method, whose frequency slowly climbs and gradually approaches the upper limit of the equipment's rated output frequency. The black solid line marked with a square represents the collaborative control method of this invention; this curve shows that the actuator rapidly exceeds the rated frequency in the initial startup phase, i.e. Figure 3In section (2), the black dotted line represents the upper limit of the rated output frequency of the 50Hz device. It is raised to the upper limit frequency of 55Hz and maintained for a short time to establish a blocking air curtain. Then it smoothly drops back to the steady-state operating frequency of 45Hz, and gradually drops back to the normal operating speed of 15Hz after the ventilation effect meets the reset conditions. The above frequency evolution reflects the control characteristics of limited high-slope start-up, short-time pressure holding, steady-state operation and smooth reset.

Claims

1. A method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model, characterized in that, The method includes: Acquire environmental status data and traffic flow video images, generate a steady-state sequence data stream containing a steady-state concentration sequence based on the environmental status data, and generate an early warning trigger signal based on the steady-state sequence data stream; In response to the warning trigger signal, vehicle motion characteristic parameters are determined based on the traffic flow video image, and multiple wind speed components of the tunnel are determined based on the environmental state data and the vehicle motion characteristic parameters; An effective driving airflow is generated based on the multi-speed components, and the diffusion mileage range of pollutants is determined based on the effective driving airflow. Interception fans and auxiliary fans are then determined within the diffusion mileage range. The steady-state concentration sequence is analyzed to generate the concentration rise rate and concentration change acceleration. The pre-scavenging air volume ratio coefficient is dynamically adjusted based on the concentration rise rate and concentration change acceleration to obtain the adjusted pre-scavenging air volume ratio coefficient. The target air volume command is generated based on the adjusted pre-scavenging air volume ratio coefficient. The tunnel hydraulic parameters and temperature gradient data are obtained, a critical anti-backflow wind speed is generated based on the tunnel hydraulic parameters and temperature gradient data, and a limited high-slope start command is generated based on the critical anti-backflow wind speed. Based on the limited high-slope start command, the interceptor fan and the auxiliary fan are controlled to perform corresponding control actions.

2. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 1, characterized in that, The environmental state data includes carbon monoxide concentration data and visibility data. The step of generating a steady-state sequence data stream containing a steady-state concentration sequence based on the environmental state data, and generating an early warning trigger signal based on the steady-state sequence data stream, includes: Dynamic noise reduction processing is performed on the carbon monoxide concentration data and the visibility data based on the Kalman filter algorithm to establish a discrete state-space model. Before the covariance matrix inversion step in the discrete state space model, a pre-defined matrix non-singularity test logic is introduced to generate an optimized model. The environmental state data is processed based on the optimized model to output a steady-state concentration sequence and a steady-state visibility sequence. A first monitoring signal is generated based on the steady-state concentration sequence and the upper limit threshold of carbon monoxide concentration, and a second monitoring signal is generated based on the steady-state visibility sequence and the lower limit threshold of visibility. An early warning trigger signal is generated based on the first monitoring signal or the second monitoring signal.

3. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 1, characterized in that, The vehicle motion characteristic parameters include vehicle density, average vehicle frontal area, and average vehicle speed. The determination of multiple wind speed components in the tunnel based on the environmental state data and the vehicle motion characteristic parameters includes: Based on the environmental state data, the natural basic wind speed component and wind direction are determined, and a reference flow direction code is generated based on the natural basic wind speed component and the wind direction. The equivalent pressure difference of the traffic piston is determined based on the vehicle density, the average frontal area of ​​the vehicle, the average speed of the vehicle, and the cross-sectional area of ​​the tunnel, and the traffic piston wind component is generated based on the equivalent pressure difference of the traffic piston. Determine the real-time temperature of the warning point corresponding to the warning trigger signal at the current moment, determine the temperature difference between the real-time temperature and the absolute temperature of the normal environment, and obtain the local longitudinal slope angle of the section corresponding to the warning trigger signal. The thermo-pressure buoyancy wind speed component is generated based on the gravitational acceleration constant, the reference length of the benchmark section, the local longitudinal slope angle, and the temperature difference.

4. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 3, characterized in that, The process of generating an effective driving airflow based on the multi-velocity components, determining the diffusion mileage range of pollutants based on the effective driving airflow, and determining the intercepting fan and auxiliary fan within the diffusion mileage range includes: An effective driving airflow is generated based on the multi-wind speed components and the reference flow direction encoding; Construct a diffusion calculation model, and determine the diffusion coverage radius based on the effective driving airflow and the diffusion calculation model; The diffusion mileage range is determined based on the diffusion coverage radius; Obtain the physical location of each wind turbine in space, and determine the intercepting wind turbine and auxiliary wind turbine based on the physical location in space and the diffusion mileage range.

5. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 1, characterized in that, The step involves analyzing the steady-state concentration sequence to generate the concentration rise rate and concentration change acceleration. Based on the concentration rise rate and concentration change acceleration, the pre-scavenging gas volume ratio is dynamically adjusted to obtain the adjusted pre-scavenging gas volume ratio, including: The first derivative of the steady-state concentration sequence is processed to obtain the concentration rise rate; Perform a quadratic difference processing on the adjacent concentration rise rates to obtain the concentration change acceleration; Obtain the tunnel cross-sectional area, and determine the coefficient base mapping value based on the concentration rise rate and the tunnel cross-sectional area; The pre-scavenging air volume ratio coefficient is dynamically adjusted based on the coefficient base mapping value, the concentration rise rate, the concentration change acceleration, the change rate compensation weight coefficient, and the change acceleration compensation weight coefficient to obtain the adjusted pre-scavenging air volume ratio coefficient.

6. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 5, characterized in that, The process of generating a critical anti-backflow wind speed based on the tunnel hydraulic parameters and the temperature gradient data, and generating a limited high-slope start command based on the critical anti-backflow wind speed, includes: The smoke front area is determined based on the aforementioned warning trigger signal; Obtain the longitudinal absolute temperature difference in the smoke front region, and determine the hydraulic equivalent diameter based on the tunnel cross-sectional area; The critical anti-backflow wind speed is determined based on the anti-backflow safety margin coefficient, the gravitational acceleration constant, the hydraulic equivalent diameter, the longitudinal absolute temperature difference, and the normal ambient absolute temperature. The steady-state operating wind speed is determined based on the pre-sweeping air volume multiple coefficient, and a steady-state operating control command is generated based on the steady-state operating wind speed and the critical anti-backflow wind speed. Obtain the amplitude limiting anti-backflow start curve, and generate an amplitude limiting high slope start command based on the steady-state operation control command and the amplitude limiting anti-backflow start curve.

7. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 1, characterized in that, After generating the warning trigger signal, the method further includes: Determine whether multiple early warning trigger signals are generated in the tunnel; If so, obtain the global environmental basic wind direction, determine the associated wind turbines corresponding to the warning trigger signal, perform spatial topology sorting on the associated wind turbines based on the global environmental basic wind direction, and obtain wind turbine sorting information, which includes upstream grouping and downstream grouping; Obtain the initial smoke exhaust velocity of the upstream group, and determine the average peak propagation velocity based on the initial smoke exhaust velocity; Determine the straight-line physical spatial distance between the upstream and downstream formations, and determine the physical propagation delay time based on the average peak propagation wind speed and the straight-line physical spatial distance; Obtain the initial dynamic pressure amplitude output by the upstream group, and determine the remaining effective wind pressure based on the initial dynamic pressure amplitude and the straight-line physical space distance.

8. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 7, characterized in that, The method further includes: Get the duration of the fan startup; When the duration reaches the physical propagation delay time, the variable frequency drive execution module corresponding to the downstream group is controlled to switch to constant output torque control mode; Acquire data on the increase in lagging wind pressure, and determine the frequency reduction amplitude required to absorb the excess increase in wind pressure based on the data on the increase in lagging wind pressure; Based on the frequency reduction amplitude, the variable frequency drive execution module corresponding to the downstream group is controlled to perform a fine-tuning downward movement of the operating frequency.

9. The method for coordinated start-up control of tunnel variable frequency fans based on a pre-scavenging air model according to claim 1, characterized in that, The method further includes: After controlling the intercepting fan and the auxiliary fan to perform the corresponding control actions, the carbon monoxide concentration value and visibility value of the alarm monitoring point corresponding to the early warning trigger signal are obtained; The ventilation effect evaluation results are generated based on the carbon monoxide concentration and visibility values. If the ventilation effect evaluation result meets the preset reset condition, the operating frequency of the control fan is restored to the steady-state operating frequency; If the ventilation effect evaluation result does not meet the preset reset condition, a full-load operation command is generated. Based on the full-load operation command, the corresponding frequency converter drive execution module is controlled to perform the corresponding full-load operation.

Images