Tunnel fire high-risk target secondary disaster risk control method, system and equipment

Abstract

The application discloses a tunnel fire high-risk target secondary disaster risk control method, system and equipment, relates to the technical field of tunnel fire safety control and high-risk target secondary disaster risk control, and comprises the following steps: acquiring fire state data, determining a high-risk target and a dangerous state criterion, a passable evacuation path and a candidate control action; predicting the smoke layer height, visibility, smoke concentration and high-risk target surface heat flow time sequence of the evacuation path coverage area for each candidate control action, determining the most unfavorable predicted smoke layer stable safety margin and the most unfavorable predicted high-risk target dangerous state safety margin; determining the longitudinal wind speed upper limit according to the surface heat flow growth rate, the fire heat release rate and the distance between the fire source and the high-risk target, determining the target control action under the condition of personnel escape safety priority, controlling the longitudinal ventilation equipment and the lateral smoke exhaust port to execute, and rolling updating according to the execution feedback. The application can give consideration to personnel evacuation safety and high-risk target secondary disaster risk control.

Description

Technical Field

[0001] This invention relates to the field of tunnel fire safety control and secondary disaster risk control for high-risk targets, and particularly to methods, systems and equipment for controlling secondary disaster risks of high-risk targets in tunnel fires. Background Technology

[0002] With the development of transportation infrastructure such as highway tunnels, urban underground roads, undersea tunnels, and underground transportation corridors, the number of vehicles passing through tunnels is constantly increasing, and the types of vehicles are becoming increasingly complex. Due to the narrow and relatively enclosed space of tunnels, after a vehicle fire, heat and smoke can easily accumulate and spread longitudinally along the tunnel, causing smoke layers to sink, visibility to decrease, and the evacuation environment to deteriorate. When there are tank trucks, hazardous chemical transport vehicles, large trucks, or vehicles carrying flammable materials near or downstream of the fire source, the heat radiation, high-temperature smoke, or flames generated by the initial fire source may continue to heat these vehicles or other high-risk targets, thereby increasing the risk of secondary combustion, explosion, or other secondary disasters.

[0003] After a tunnel fire, the flow of smoke within the tunnel is typically regulated using longitudinal ventilation equipment, centralized smoke extraction systems, key smoke extraction systems, or lateral smoke extraction outlets to limit smoke spread and maintain the environmental conditions necessary for personnel evacuation. Relevant control schemes can be implemented by coordinating the adjustment of longitudinal ventilation and smoke extraction systems based on information such as the location of the fire source, tunnel temperature, longitudinal wind speed, and the operational status of the smoke extraction facilities.

[0004] In developing this invention, the inventors discovered that changes in longitudinal ventilation intensity and smoke exhaust outlet status not only affect smoke layer height, visibility, and smoke concentration, but may also alter flame morphology and the transport direction of high-temperature smoke. When ventilation and smoke exhaust actions are primarily determined based on smoke emission status or personnel evacuation conditions, insufficient consideration may be given to the heat risk to adjacent or downstream high-risk targets; conversely, limiting longitudinal ventilation intensity to reduce the continuous heat risk to high-risk targets may affect smoke control effectiveness and personnel evacuation conditions. Therefore, related control schemes struggle to simultaneously address both personnel evacuation conditions and the secondary disaster risks to high-risk targets during the same fire response process. Summary of the Invention

[0005] Based on the technical problems existing in the background technology, the present invention proposes a method, system and equipment for controlling the risk of secondary disasters of high-risk targets in tunnel fires, which can take into account both the safety of personnel evacuation and the control of secondary disaster risks of high-risk targets.

[0006] This invention proposes a method for controlling secondary disaster risks at high-risk targets in tunnel fires, comprising: Acquire fire scene status data for tunnel fires; Based on the fire situation data, identify one or more high-risk targets and the corresponding danger status criteria for each high-risk target, and determine the passable evacuation routes based on the location of trapped personnel, the current available safety exits, and the status of evacuation routes; Under the current ventilation and smoke extraction conditions, the smoke layer height, visibility, and smoke concentration in the area covered by the passable evacuation route are estimated based on the fire situation data, and the current smoke layer stability safety margin is determined; the surface heat flow sequence of each high-risk target in the prediction time domain is predicted, and the current high-risk target danger state safety margin is determined. When the current flue gas layer stability safety margin is lower than the first preset threshold, or the current high-risk target danger state safety margin is lower than the second preset threshold, a set of candidate control actions is constructed. Based on each candidate control action, predict the smoke layer height, visibility, and smoke concentration in the coverage area of ​​the passable evacuation path within the prediction time domain, and predict the surface heat flow sequence of each high-risk target within the prediction time domain; determine the smoke layer stability safety margin sequence based on the smoke layer height, visibility, and smoke concentration corresponding to each prediction time, and determine the minimum value in the smoke layer stability safety margin sequence as the most unfavorable predicted smoke layer stability safety margin for the corresponding candidate control action; determine the remaining evacuation time of personnel based on the personnel location, exit status, and path status corresponding to each prediction time under the corresponding candidate control action; determine the remaining arrival time of danger for each high-risk target from the corresponding prediction time based on the surface heat flow sequence and corresponding hazard status criteria; determine the hazard status safety margin sequence of high-risk targets based on the remaining arrival time of danger, the remaining evacuation time of personnel, and the preset safety redundancy time, and determine the minimum value in the hazard status safety margin sequence of high-risk targets as the most unfavorable predicted hazard status safety margin for the corresponding candidate control action; Based on the surface heat flow growth rate of each high-risk target, the roof temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target, the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined, and the minimum value among the upper limits of the longitudinal wind speed corresponding to each high-risk target is determined as the global upper limit of the longitudinal wind speed. From the candidate control actions that meet the global longitudinal wind speed upper limit requirements, select the candidate control actions that have a safety margin of flue gas layer stability of the most unfavorable prediction not lower than the first preset threshold and a safety margin of dangerous state of the most unfavorable prediction not lower than the second preset threshold, and determine the target control action based on the selection results; when there are no candidate control actions that simultaneously meet the first preset threshold and the second preset threshold, determine the target control action in the order of priority of meeting personnel escape safety constraints over meeting secondary disaster risk control constraints of high-risk targets. Control the longitudinal ventilation equipment and lateral smoke exhaust outlets to execute target control actions, collect actual longitudinal wind speed, longitudinal ventilation equipment operating status and lateral smoke exhaust outlet positioning status, update fire scene status data, available equipment status and candidate control action set based on the collected results, and re-predict and evaluate candidate control actions and determine target control actions; When the current smoke layer stability safety margin is not lower than the first preset threshold and the current high-risk target danger state safety margin is not lower than the second preset threshold, the current ventilation and smoke exhaust state is maintained and fire status data is continuously acquired.

[0007] Preferably, the fire situation data includes the roof temperature field, heat flow of the target or adjacent area, longitudinal wind speed, status of lateral smoke vents, vehicle identification results, fire source location and fire source heat release rate, location and evacuation status of trapped personnel, status of currently available safety exits, status of evacuation routes, and tunnel geometry and topology data. The fire scene status data is processed by timestamp alignment, spatial coordinate mapping, outlier removal, missing value compensation, and filtering. The tunnel space is divided into multiple control units along the longitudinal direction, and each control unit is further divided into personnel evacuation layer and smoke layer along the height direction. Fire sources, high-risk targets, trapped personnel, safety exits, temperature measuring points, heat flow measuring points, lateral smoke exhaust outlets, and execution equipment are mapped to the corresponding control units according to the longitudinal coordinates, the height layer they are located in, and the boundary of the control unit. For an object located at the boundary of an adjacent control unit, map it to a control unit that is closer to the object; when the spatial influence range of the same object covers multiple control units, establish the association between the object and the corresponding control unit respectively.

[0008] Preferably, the vehicle identification results include vehicle type, spatial location, operating status, stationary status, and identification confidence level; Based on the vehicle identification results, the target type, spatial location, and identification confidence level of the high-risk target are determined, and the heat exposure direction of the high-risk target is determined based on the spatial relationship between the fire source and the high-risk target. Obtain the equivalent size, equivalent heated area, hazard level, hazard parameters, threshold source and hazard status criterion corresponding to the target type from the target type-hazard parameter rule table, and convert high-risk targets into potential secondary ignition sources or hazard status targets. Hazard parameters include critical surface heat flux, critical cumulative heat dose and critical surface temperature. When the identification confidence level is lower than the preset identification confidence level threshold, the high-risk target is corrected to a vehicle type of the same size level or higher risk level, and the risk parameters and risk status criteria corresponding to the corrected vehicle type are adopted.

[0009] Preferably, for each control unit covered by the passable evacuation path, a sub-margin of smoke layer height is determined based on the normalized deviation of smoke layer height relative to the minimum safe smoke layer height, a sub-margin of visibility is determined based on the normalized deviation of visibility relative to the minimum safe visibility, and a sub-margin of smoke concentration is determined based on the normalized deviation of smoke concentration relative to the safe threshold of smoke concentration. The minimum value among the flue gas layer height sub-margin, visibility sub-margin, and flue gas concentration sub-margin is determined as the flue gas layer stability safety margin of the corresponding control unit, and the minimum value among the flue gas layer stability safety margins corresponding to each prediction time and each control unit in the prediction time domain is determined as the most unfavorable predicted flue gas layer stability safety margin. The auxiliary stability ranking index is determined based on the ceiling temperature gradient and longitudinal wind speed disturbance between adjacent control units. The auxiliary stability ranking index is used to rank the candidate control actions that meet the personnel escape safety constraints when determining the target control action.

[0010] Preferably, the surface heat flow time series is determined based on a pre-established heat flow time series database, which is indexed by the heat release rate of the fire source, the longitudinal wind speed, the distance between the fire source and the high-risk target, the height layer where the high-risk target is located, the status of the lateral smoke exhaust outlet, and the scale attributes of the high-risk target. When the heat release rate of the fire source, longitudinal wind speed, distance between the fire source and the high-risk target, height level of the high-risk target, status of the lateral smoke exhaust outlet, and scale attributes of the current fire condition are within the coverage of the typical fire condition target surface heat flow time series database, the baseline surface heat flow time series is obtained by nearest neighbor condition matching or multilinear interpolation. When any current operating condition parameter exceeds the coverage of the typical operating condition target surface heat flow time series database, the surface heat flow time series corresponding to the conservative boundary condition that is closest to the current fire condition is adopted. Based on the target type and heat exposure direction of high-risk targets, target type correction coefficients and heat exposure direction correction coefficients are determined to correct the baseline surface heat flow time series. Based on the changes in roof temperature, roof temperature gradient, longitudinal wind speed, lateral smoke exhaust port opening and closing, and historical heat flow data of the target surface in the control unit where the corresponding high-risk target is located, residual correction is performed on the corrected surface heat flow time series, and the residual correction amount is continuously updated according to the deviation between the latest measured heat flow and the predicted heat flow.

[0011] Preferably, the cumulative heat dose of the corresponding high-risk target in the prediction time domain is determined based on the surface heat flow time series and sampling time interval of each high-risk target; The hazard condition criteria include one or more of the following: the target surface heat flux reaches or exceeds the critical surface heat flux, the cumulative heat dose reaches or exceeds the critical cumulative heat dose, and the target surface temperature reaches or exceeds the critical surface temperature. The time between the current evaluation time and the earliest predicted time that the hazard condition criteria are met is determined as the hazard arrival time. For each trapped person, paths and safety exits that are blocked by fire sources, have smoke environments that do not meet the passage conditions, or are closed are removed from the tunnel topology map. Based on the path length, the evacuation speed of the person and the congestion status of the path, the evacuation time of the trapped person to each currently available safety exit is determined, and the minimum value among them is determined as the evacuation time of the trapped person. The maximum evacuation time for all trapped personnel is determined by the sum of the personnel reaction time and the congestion correction time. The difference between the remaining time of danger arrival and the time of personnel evacuation and the preset safety redundancy time is determined as the safety margin of the corresponding high-risk target's dangerous state. If the dangerous state is not reached within the prediction time domain, the remaining time for the corresponding danger to arrive is recorded as the right censoring quantity. The lower bound of the safety margin of the dangerous state is determined based on the duration from the current evaluation time to the end of the prediction time domain, the personnel evacuation time, and the preset safety redundancy time. Only when the lower bound of the safety margin of the dangerous state is not lower than the second preset threshold is the corresponding candidate control action determined to meet the hard constraint of the dangerous state of the high-risk target.

[0012] Preferably, the upper limit of longitudinal wind speed for each high-risk target is determined from the wind speed limiting function calibrated by fire test or fire simulation based on the surface heat flow growth rate of each high-risk target, the ceiling temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target. The minimum value among the upper limits of longitudinal wind speed for each high-risk target is determined as the global upper limit of longitudinal wind speed. When the longitudinal wind speed has a non-monotonic relationship with the dangerous state of the high-risk target, the minimum longitudinal wind speed at which the unacceptable dangerous state first appears is used as the boundary to remove candidate control actions from the candidate control action set whose candidate longitudinal wind speed is not lower than the minimum longitudinal wind speed. When the candidate control actions that simultaneously satisfy the first preset threshold and the second preset threshold are not empty, the candidate control actions that do not reach the dangerous state of each high-risk target in the prediction time domain and whose lower limit of the safety margin of the dangerous state satisfies the second preset threshold are selected first. Among the candidate control actions with the same right censoring category, the target control action is determined according to the overall level of surface heat flow of the high-risk target in the prediction time domain, the predicted terminal heat flow or cumulative heat dose, and the change in control action. When there are no candidate control actions that simultaneously meet the first and second preset thresholds, the candidate control actions that meet the global longitudinal wind speed upper limit requirements are sorted in order of the degree of satisfaction of personnel escape safety constraints, the safety margin of the most unfavorable predicted high-risk target dangerous state, the predicted terminal heat flow or cumulative heat dose, the surface heat flow growth rate and the change in control action, and the target control action is determined based on the sorting results. When the execution feedback indicates that the deviation between the actual longitudinal wind speed and the target longitudinal wind speed exceeds the preset tolerance, the longitudinal ventilation equipment has not reached the target operating state, the lateral smoke exhaust port has not reached the target opening and closing position within the preset time, the actuator has malfunctioned, or the corresponding action is not executable, the candidate control action containing the corresponding faulty equipment state or unexecutable state is removed from the candidate control action set, and the target control action is re-determined.

[0013] This invention proposes a secondary disaster risk control system for high-risk targets in tunnel fires, comprising: A multi-source data acquisition module is used to acquire fire scene status data in tunnel fires; The two-layer spatial modeling module is used to divide the tunnel space into a personnel evacuation layer and a smoke layer, and further divide it into multiple control units along the longitudinal direction of the tunnel. Based on the location of the trapped personnel, the currently available safety exits, and the status of the evacuation routes, the module determines the passable evacuation routes. The high-risk target equivalent module is used to determine one or more high-risk targets, the control unit where each high-risk target is located, and the danger status criteria corresponding to each high-risk target based on fire scene status data. The smoke layer stability assessment module is used to estimate the smoke layer height, visibility, and smoke concentration in the area covered by the passable evacuation path under the current ventilation and smoke exhaust conditions, based on fire state data. It also determines the current smoke layer stability safety margin based on the margins of the smoke layer height, visibility, and smoke concentration relative to the corresponding safety thresholds. Furthermore, it is used to predict the smoke state for each candidate control action, determine the smoke layer stability safety margin sequence based on the predicted smoke state, and determine the minimum value in the smoke layer stability safety margin sequence as the most unfavorable predicted smoke layer stability safety margin for the corresponding candidate control action. The heat flow prediction and risk module is used to predict the surface heat flow sequence of each high-risk target in the prediction time domain. Based on the surface heat flow sequence and the corresponding hazard status criteria, it determines the remaining time of arrival of the hazard for each high-risk target, determines the personnel evacuation time based on the passable evacuation path, and determines the safety margin of the current high-risk target's hazard status based on the difference between the remaining time of arrival of the hazard and the personnel evacuation time and the preset safety redundancy time. It is also used to predict the surface heat flow sequence of high-risk targets for each candidate control action, determine the safety margin sequence of the high-risk target's hazard status based on the remaining personnel evacuation time, the remaining time of arrival of the hazard, and the preset safety redundancy time, and determine the minimum value in the high-risk target's hazard status safety margin sequence as the most unfavorable predicted high-risk target hazard status safety margin for the corresponding candidate control action. The wind-assisted fire spread limiting and control decision module is used to construct a set of candidate control actions when the current smoke layer stability safety margin is lower than a first preset threshold or the current high-risk target danger state safety margin is lower than a second preset threshold. Each candidate control action includes a candidate longitudinal wind speed and a candidate combination of opening and closing lateral smoke exhaust ports. Based on the surface heat flux growth rate of each high-risk target, the roof temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target, the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined, and the minimum value among the upper limits of the longitudinal wind speed corresponding to each target is determined as the global upper limit of the longitudinal wind speed. The module then selects the target that meets the global upper limit of the longitudinal wind speed. Among the required candidate control actions, those with the most unfavorable predicted flue gas layer stability safety margin not lower than the first preset threshold and the most unfavorable predicted high-risk target danger state safety margin not lower than the second preset threshold are selected, and the target control action is determined based on the selection results; when there are no candidate control actions that simultaneously meet the first and second preset thresholds, the target control action is determined in the order of priority of personnel escape safety constraints over high-risk target secondary disaster risk control constraints; when the current flue gas layer stability safety margin is not lower than the first preset threshold and the current high-risk target danger state safety margin is not lower than the second preset threshold, the current ventilation and smoke exhaust state is maintained. The execution and feedback module is used to control the longitudinal ventilation equipment and lateral smoke exhaust outlets to perform target control actions, collect actual longitudinal wind speed, longitudinal ventilation equipment operating status, and lateral smoke exhaust outlet positioning status, and feed the collected results back to the multi-source data acquisition module and the wind-assisted fire intensity limiting and control decision module to update fire scene status data, available equipment status, and candidate control action set, and to re-predict and evaluate candidate control actions and determine target control actions; it is also used to control the multi-source data acquisition module to continuously acquire fire scene status data while maintaining the current ventilation and smoke exhaust status.

[0014] The present invention provides an electronic device comprising a processor, a memory, a data interface, and a control interface; The memory is connected to the processor and is used to store control programs that can be executed by the processor, tunnel geometry and topology data, target type parameter tables, and heat flow timing databases; The data interface is used to receive timestamped fire status data, trapped personnel data, safety exit data, and evacuation route data; The control interface is used to connect the controller of the longitudinal ventilation equipment and the actuator of the lateral smoke exhaust outlet, send the target longitudinal wind speed and the opening and closing commands of the lateral smoke exhaust outlet, and receive the execution status feedback; The processor is used to execute control programs to enable electronic devices to implement the aforementioned secondary disaster risk control methods for high-risk targets in tunnel fires.

[0015] The present invention proposes a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the above-mentioned method for controlling secondary disaster risks of high-risk targets in tunnel fires.

[0016] Beneficial technical effects of the present invention: (1) This invention simultaneously predicts the smoke state and heat flow sequence of the surface of high-risk targets in the area covered by the passable evacuation path for the same candidate longitudinal wind speed and candidate lateral smoke outlet opening and closing combination. The most unfavorable smoke layer stability safety margin and the most unfavorable high-risk target danger state safety margin in the prediction time domain are used as the screening basis for candidate control actions, so that the control decision can simultaneously reflect the changes in the personnel evacuation environment and the heating state of high-risk targets, avoid the risks that occur in local areas or at specific prediction times being covered by other safety states, and improve the adaptability of target control actions to the most unfavorable fire state in the prediction time domain.

[0017] (2) This invention compares the remaining time for the danger to reach a high-risk target with the time for personnel evacuation and the preset safety redundancy time to form a safety margin for the dangerous state of the high-risk target. It transforms the possibility of a high-risk target reaching a dangerous state before personnel evacuation is completed into a quantifiable time safety relationship, thereby providing a clear temporal basis for the risk assessment of secondary disasters under different candidate control actions. This is conducive to reducing the risk of high-risk targets forming secondary fire sources, causing explosions, or entering other dangerous states before personnel evacuation is completed.

[0018] (3) Based on the surface heat flow growth rate of the high-risk target, the ceiling temperature gradient of the corresponding control unit, the heat release rate of the fire source, and the distance between the fire source and the high-risk target, the present invention determines the upper limit of the longitudinal wind speed corresponding to each high-risk target, and takes the minimum value as the global upper limit of the longitudinal wind speed, so that the selection of candidate control actions is limited by the actual heat evolution state of the high-risk target, which is conducive to suppressing the flame stretching, hot smoke shift and heat enhancement of the high-risk target caused by excessive longitudinal wind speed, and taking into account the smoke control needs of the personnel evacuation area through the combination of longitudinal ventilation equipment and lateral smoke exhaust outlet.

[0019] (4) When there are no candidate control actions that simultaneously satisfy both types of safety constraints, the present invention determines the target control action in the order of the degree of satisfaction of personnel escape safety constraints taking precedence over the degree of satisfaction of secondary disaster risk control constraints of high-risk targets. This ensures that when the two safety constraints conflict, a clear control priority can still be maintained, guaranteeing that the control action first serves the safe evacuation of trapped personnel, while reducing the risk of secondary disasters of high-risk targets within the global longitudinal wind speed limit.

[0020] (5) The present invention collects the actual longitudinal wind speed, the operating status of the longitudinal ventilation equipment and the position status of the lateral smoke exhaust outlet, and updates the fire status data, available equipment status and candidate control action set according to the collection results, so that the system can re-predict and evaluate the candidate control actions and determine the target control actions according to the fire development and equipment execution status, reduce the deviation between the predicted control actions and the actual execution status, and improve the continuity and closed-loop regulation reliability of the tunnel fire risk control process. Attached Figure Description

[0021] Figure 1 This is a flowchart of the secondary disaster risk control method for high-risk targets in tunnel fires proposed in this invention; Figure 2 This is a block diagram showing the composition and data closed-loop of the secondary disaster risk control system for high-risk targets in tunnel fires proposed in this invention. Figure 3 This is a schematic diagram showing the arrangement of the two-layer tunnel space, the longitudinal control unit, and the sensing and actuation equipment proposed in this invention. Detailed Implementation

[0022] The present invention will be further explained below with reference to specific embodiments.

[0023] Example 1 like Figure 1 As shown, this embodiment provides a method for controlling the secondary disaster risk of high-risk targets in tunnel fires. This method acquires and processes fire scene data from tunnel fires, establishes a two-layer tunnel space and a longitudinal control unit, performs conservative equivalence on high-risk targets, evaluates the smoke state in personnel evacuation areas and the heat risk of high-risk targets, and uses the smoke layer stability safety margin and the high-risk target danger state safety margin as the basis for judging the coordinated control of longitudinal ventilation and lateral smoke exhaust.

[0024] When the current flue gas layer stability safety margin is lower than the first preset threshold, or the current high-risk target danger state safety margin is lower than the second preset threshold, the candidate longitudinal wind speed and candidate lateral smoke exhaust port opening and closing combination are predicted and evaluated. Under the longitudinal wind speed limit condition, the target control action is determined and rolled adjustment is made according to the equipment execution feedback and the updated fire scene status.

[0025] First, acquire fire scene status data of the tunnel fire and form a fire scene status dataset Y(t). The fire scene status data includes at least the ceiling temperature field, heat flow of the target or adjacent area, longitudinal wind speed, status of lateral smoke exhaust outlets, vehicle identification results, fire source location and heat release rate, location and evacuation status of trapped personnel, status of currently available safety exits, status of evacuation routes, and tunnel geometry and topology data.

[0026] The fire scene status dataset is represented as follows: (1) In the formula, t is the current sampling time; T c (t) represents the roof temperature field; q(t) represents the heat flux data of the high-risk target surface or adjacent area; u(t) represents the longitudinal wind speed; S e (t) represents the state of the side exhaust port; R v (t) represents the vehicle identification result; F(t) represents the location of the fire source, the heat release rate of the fire source, and the stage of fire development; P(t) represents the location, number, and evacuation status of the trapped personnel; E(t) represents the status of currently available safe exits and evacuation routes; G represents the tunnel geometry and topology data.

[0027] The ceiling temperature field can be obtained by ceiling temperature sensors deployed longitudinally along the tunnel. Heat flow on the target surface or adjacent areas can be obtained by heat flow measurement points located near key monitoring areas or high-risk targets. If no heat flow measurement points are installed on the surface of a high-risk target, heat flow measurement points in adjacent areas within the same control unit, heat flow measurement points in adjacent control units, calculated ceiling temperature field values, or corresponding predicted values ​​from a heat flow time-series database can be used as substitute data.

[0028] Longitudinal wind speed is obtained from a wind speed sensor. The status of the lateral exhaust vents includes at least their open, closed, and in-position states. Vehicle identification results are obtained from a video recognition system or traffic monitoring system and include at least vehicle type, spatial location, operating status, stationary status, and identification confidence level.

[0029] The location of the fire source and the heat release rate can be obtained from at least one of the following: fire detection system, video recognition results, and heat release rate inversion model. The location, number, evacuation status, and congestion status of trapped personnel can be obtained from video recognition system, traffic monitoring system, or evacuation guidance system. The location and current availability of safety exits can be obtained from a pre-set tunnel topology database and escape door monitoring system.

[0030] Because the data sources differ, the sampling times and coordinate references for various data types may vary. Therefore, the fire situation data undergoes timestamp alignment and spatial coordinate mapping to ensure that different data correspond to a unified evaluation time and tunnel longitudinal coordinate system. Outliers in the fire situation data are removed, missing data is compensated for, and data with measurement fluctuations are filtered.

[0031] Outliers can be identified based on sensor range, data change rate, or differences between adjacent measurement points. Missing value compensation can be achieved through interpolation between adjacent sampling times, substitution with data from nearby measurement points, or by using the corresponding prediction model output value. Filtering can employ moving average filtering, exponential filtering, or other filtering methods that suppress random noise while preserving the trends in temperature, heat flux, and wind speed changes.

[0032] like Figure 3 As shown, the tunnel space is divided longitudinally into multiple control units, and each control unit is further divided vertically into a personnel evacuation layer and a smoke layer. The personnel evacuation layer is located between the tunnel surface and the preset safe height for personnel, and is used to characterize the evacuation status of trapped personnel, the location of high-risk targets, and the thermal state of high-risk targets. The smoke layer is located above the personnel evacuation layer and between the tunnel ceiling, and is used to characterize the smoke layer height, visibility, smoke concentration, ceiling temperature, and smoke disturbance state.

[0033] The j-th control unit can be represented as: (2) in, For the j-th tunnel control unit, For the personnel evacuation layer in the j-th control unit, Let J be the flue gas layer in the j-th control unit.

[0034] The longitudinal length of the control unit can be determined based on the spacing between ceiling temperature measuring points, the spacing between lateral smoke exhaust outlets, the spacing between heat flow measuring points, and the control accuracy. The division of the control unit should remain consistent within the same control process to ensure that the state data at each sampling and prediction time have a unified spatial reference.

[0035] The fire source, high-risk targets, trapped personnel, safety exits, temperature measuring points, heat flow measuring points, lateral smoke exhaust vents, and actuators are mapped to the corresponding control units according to their longitudinal coordinates, the height level they are located in, and the control unit boundaries. For objects located at the boundaries of adjacent control units, the object is mapped to the control unit that is closer to it. This distance can be the distance between the object's longitudinal coordinate and the center position of the adjacent control unit.

[0036] When the spatial influence range of the same object covers multiple control units, the association between the object and the corresponding control unit is established separately. For example, when the thermal influence range of a fire source covers the control unit where the fire source is located and its downstream adjacent control units, the fire source status is simultaneously associated with the aforementioned control units; when the length of a high-risk target spans two control units, the high-risk target is associated with each of the two control units separately.

[0037] Based on the location of the trapped personnel, the status of currently available safety exits, and the evacuation routes, a passable evacuation route is determined. Specifically, in the tunnel topology map, routes and safety exits that are blocked by fire sources, whose smoke environment does not meet the passage conditions, or that are closed are eliminated. Then, the passable evacuation routes are determined based on the connectivity between the current location of the trapped personnel and each currently available safety exit.

[0038] Control units covered by accessible evacuation routes are used to determine the evaluation range of flue gas stability safety margin. When there are multiple trapped personnel, each trapped personnel can be associated with one or more accessible evacuation routes, and the union of control units covered by all accessible evacuation routes is taken as the flue gas condition evaluation area.

[0039] One or more high-risk targets are identified based on the vehicle identification results. High-risk targets may include ordinary passenger cars, light passenger vehicles, large passenger buses, ordinary trucks, tank trucks, flammable liquid transport vehicles, liquefied petroleum gas transport vehicles, compressed natural gas transport vehicles, hydrogen pressure vessel vehicles, hazardous chemical transport vehicles, vehicles carrying combustibles, and other targets that may form potential secondary ignition sources or enter a dangerous state under the action of an initial ignition source.

[0040] The target type, spatial location, and identification confidence level of high-risk targets are determined based on vehicle type, spatial location, and identification confidence level. The direction of heat exposure is then determined based on the spatial relationship between the fire source and the high-risk target. The direction of heat exposure can include lateral heating, frontal heating, rear-end heating, or oblique heating.

[0041] High-risk targets are conservatively equivalentd using a target type-hazard parameter rule table. The conservative equivalent state of the i-th high-risk target is represented as: (3) In the formula, For the first Equivalent vehicle type for a high-risk target For spatial location, , , These are equivalent length, equivalent width, and equivalent height, respectively. For equivalent heating area, As a weight for the risk level, The critical heat flux threshold, The critical cumulative heat dose threshold, This is the critical surface temperature threshold.

[0042] Obtain the equivalent size, equivalent heated area, hazard level, hazard parameters, threshold source, and hazard status criteria corresponding to the target type from the target type-hazard parameter rule table. Hazard parameters include at least the critical surface heat flux, critical cumulative heat flux, and critical surface temperature.

[0043] In this embodiment, the target type-hazard parameter rule table can use the initialization parameters shown in Table 1.

[0044] Table 1 Target Type - Hazard Parameter Rule Table The parameters in Table 1 are used for system initialization and implementation examples. In actual use, the hazard parameters of the corresponding target can be determined based on physical tests, vehicle files, container type tests, project calibration results, or safety data sheets. The hazard level is used to select the corresponding hazard parameters, safety redundancy, and risk priority, and does not directly change the predicted external physical heat flux value.

[0045] Equivalent dimensions can include equivalent length, equivalent width, and equivalent height. The equivalent heated area is determined based on the target equivalent dimensions and the direction of heat exposure. When using surface heat flux per unit area as the hazard criterion, the equivalent heated area is not directly multiplied by the surface heat flux per unit area; when using total heat flux or total heat dose as the hazard criterion, the corresponding total is determined based on the surface heat flux per unit area and the equivalent heated area.

[0046] When the identification confidence level is lower than the preset identification confidence level threshold, the high-risk target is corrected to a vehicle type of the same size level or a higher risk level, and the risk parameters and risk status criteria corresponding to the corrected vehicle type are adopted to avoid underestimating the risk of high-risk targets due to uncertainty in vehicle type identification.

[0047] Under the current ventilation and smoke extraction conditions, the smoke layer height, visibility, and smoke concentration of each control unit within the area covered by the passable evacuation route are estimated based on fire situation data. Smoke concentration can be selected as an evaluation index from carbon monoxide concentration, particulate matter concentration, or extinction coefficient, and this index should be kept consistent throughout the same control process.

[0048] For the j-th control unit, the flue gas state vector can be represented as: (4) In the formula, , and These represent the smoke layer height, smoke concentration, and visibility of the j-th control unit at time k, respectively.

[0049] The flue gas condition can be estimated based on the historical temperature sequence of the roof, longitudinal wind speed, the condition of the lateral smoke vents, and tunnel geometry parameters: (5) In the formula, For the estimated flue gas layer height of the j-th control unit, To estimate flue gas concentration, To estimate visibility, For the j-th control unit, the ceiling thermocouple temperature sequence within the historical time window. The current longitudinal wind speed, In order to be with the first The opening and closing status of the side exhaust vents associated with each control unit, wherein , The state vector from the global side exhaust port Extract the mapping operator associated with the state of the j-th control unit. For tunnel geometric parameters, This is the flue gas layer state estimation function.

[0050] Flue gas state estimation function Calibration can be achieved using typical tunnel fire test data, validated numerical simulation data, or historical engineering data. Alternatively, it can be implemented using at least one of the following methods: calibration database lookup and multilinear interpolation, piecewise regression, temperature profile threshold identification, state transition model, recursive regression model, or reduced-order model.

[0051] In one embodiment, the flue gas state is predicted using a first-order state recursion method: (6) In the formula, p is the candidate control action number; The state of the flue gas in the j-th control unit at the k-th prediction time under the p-th candidate control action; This includes ceiling temperature, temperature gradient, heat release rate from the fire source, candidate longitudinal wind speed, and candidate smoke exhaust outlet status; The flue gas status of the adjacent control unit; Used to limit the results to the physical range of flue gas height, concentration, and visibility; A j B j D j and c j Calibrate using historical experimental or engineering data.

[0052] The physical range constraint function is used to ensure that the smoke layer height does not exceed the tunnel clearance, the smoke concentration is not lower than zero, and the visibility is within the effective range calibrated by the model. Using the current measured or estimated smoke state as the initial state, and substituting the current longitudinal wind speed and the current smoke exhaust outlet state into the above recursive model, the smoke state under the current ventilation and smoke exhaust conditions can be obtained.

[0053] For each control unit covered by a passable evacuation route, the normalized sub-margin is determined based on the smoke layer height, visibility, and smoke concentration: (7) (8) (9) In the formula, h safe Minimum safe flue gas height; V safe For minimum safe visibility; C safe The safe threshold for flue gas concentration; , and These are the corresponding normalized sub-margins.

[0054] The minimum value among the three normalized sub-margins is determined as the flue gas layer stability safety margin for the j-th control unit: (10) In the formula, This represents the safety margin for flue gas layer stability in the j-th control unit.

[0055] When any normalization margin is lower than the corresponding safety limit, the control unit is determined to fail to meet the personnel escape safety requirements, and the excess margin of other indicators is not used to compensate for the excess indicator.

[0056] The set of control units covered by the accessible evacuation routes is denoted as . The current safety margin for flue gas layer stability is: (11) In the formula, To ensure the overall safety margin for flue gas layer stability; This is the set of control units corresponding to the personnel evacuation routes.

[0057] By taking the minimum value of each control unit within the evacuation route coverage area, the control unit that causes the most significant decrease in smoke layer, visibility, or smoke concentration is used as the basis for current personnel escape safety assessment.

[0058] The ceiling temperature gradient is determined based on the average ceiling temperature between adjacent control units. Control units located at the beginning and end of the tunnel use a one-sided differential method, while control units in the middle of the tunnel use an adjacent control unit differential method. The ceiling temperature gradient of the middle control unit can be expressed as: (12) In the formula, Let the ceiling temperature gradient be the j-th control unit. and These are the average ceiling temperatures of adjacent control units. and These represent the center positions of adjacent control units.

[0059] The longitudinal wind speed disturbance can be expressed as: (13) Auxiliary stability ranking indices were determined based on the roof temperature gradient and longitudinal wind speed disturbance. (14) In the formula, and These represent the auxiliary stability sub-indices formed by the roof temperature gradient and longitudinal wind speed disturbance, respectively. and The ranking weights are non-negative. The auxiliary stability ranking index is only used to rank candidate control actions that have met the hard constraints of personnel escape safety, and is not used to compensate for situations where the smoke layer height, visibility, or smoke concentration does not meet the corresponding safety thresholds.

[0060] The surface heat flux time series of each high-risk target is predicted within the prediction time domain. The surface heat flux time series is determined based on a pre-established heat flux time series database. This heat flux time series database can be called the typical operating condition target surface heat flux time series database, and can be represented as: (15) In the formula, This is a time-series database of heat flux on the target surface under typical operating conditions. For the heat release rate of a type a fire source, For category b longitudinal wind speed, For the c-th type of fire source-target distance, For the l-th height layer, This represents the typical open state of the m-th type of side exhaust vent. For time-series data showing the change of heat flux on the target surface over time under corresponding operating conditions, s n It represents the scale attribute of the nth type of high-risk target and corresponds one-to-one with the target surface heat flow time series in the database.

[0061] When the current fire condition's heat release rate, longitudinal wind speed, distance between the fire source and high-risk targets, height level of the high-risk targets, status of lateral smoke vents, and scale attributes are within the coverage of the heat flow time series database, the baseline surface heat flow time series is obtained through nearest neighbor condition matching or multilinear interpolation: (16) In the formula, For database condition matching or interpolation functions; To predict the length of the time domain; For the first Baseline surface heat flux time series of a high-risk target Let be the scale attribute of the i-th high-risk target.

[0062] When any current operating condition parameter exceeds the coverage of the heat flow time series database, the surface heat flow time series corresponding to the conservative boundary condition closest to the current fire condition is adopted, and parameters outside the database range are not extrapolated without constraints.

[0063] Determine the target type correction factor based on the target type and heat exposure direction of the high-risk target. and heating direction correction factor Correcting the baseline surface heat flow timing: (17) In the formula, To predict the number of discrete time steps in the time domain, This is the predicted surface heat flux per unit area after correction for target type and heating direction.

[0064] Further residual corrections are performed based on the changes in roof temperature, roof temperature gradient, longitudinal wind speed, lateral exhaust vent opening and closing, and historical heat flow data of the target surface within the control unit where the corresponding high-risk target is located. The residual correction amount can be expressed as: (18) In the formula, For the first The control unit number where the high-risk target is located. This refers to the change in ceiling temperature for the control unit. This refers to the change in the opening and closing of the lateral smoke exhaust vent associated with the control unit. For the target surface heat flow history sequence, This is the heat flux residual correction function.

[0065] In one implementation, the residual is first determined based on the deviation between the latest measured heat flow and the baseline predicted heat flow: (19) In the formula, The heat flow prediction residual at the k-th sampling time; Let be the measured surface heat flow of the i-th high-risk target at the k-th sampling time; The baseline predicted surface heat flow is given at the same time.

[0066] Update the residuals using an exponentially weighted method: (20) In the formula, For the residual update coefficients, and ; This is the residual correction amount after exponential weighting update up to the k-th sampling time.

[0067] The predicted surface heat flow in the time domain can be expressed as: (twenty one) In the formula, Baseline heat flux obtained by interpolation from the database; This is the residual attenuation coefficient. , and These are correction coefficients for changes in ceiling temperature, longitudinal wind speed, and exhaust outlet status. The relevant model coefficients need to be calibrated using experimental, validated numerical simulation, or historical engineering data that includes smoke layer height, visibility, smoke concentration, longitudinal wind speed, and exhaust outlet status.

[0068] Based on the surface heat flux time series and sampling time interval of the i-th high-risk target, determine the cumulative heat dose of the high-risk target in the prediction time domain: (twenty two) in, For the first A high-risk target Cumulative heat dose at time, The accumulated heat dose at the current moment, This represents the sampling time interval.

[0069] The criteria for determining a dangerous state include one or more of the following: (twenty three) (twenty four) (25) In the formula, For the first Critical heat flux threshold for a high-risk target The critical cumulative heat dose threshold, The critical surface temperature threshold. For the first The predicted surface temperature of a high-risk target can be determined by the surface heat flow time series and cumulative heat dose based on the heat flow-temperature response relationship calibrated by experiments, numerical simulations or engineering databases.

[0070] For tank trucks, flammable liquid transport vehicles, pressure vessel vehicles, or hazardous chemical transport vehicles, the safety valve status, internal pressure, tank wall design temperature, or specific criteria for the corresponding medium can also be used as criteria for determining hazardous conditions.

[0071] The time elapsed from the current assessment time to the earliest predicted time that any hazard criterion is met is defined as the remaining time to hazard arrival: (26) The evacuation time for the trapped person is determined based on the available evacuation routes. For the g-th trapped person, the evacuation time to reach each currently available safe exit via each available evacuation route is determined based on the route length, evacuation speed, and route congestion status. The minimum value among them is determined as the evacuation time for the trapped person: (27) In the formula, For the first A list of the currently accessible safe exits for the trapped individuals; This indicates the minimum evacuation time for the person to reach each currently available safe exit via a passable route.

[0072] The maximum evacuation time for all trapped personnel is determined by the sum of the personnel reaction time and the congestion correction time. (28) In the formula, To assemble the trapped personnel; For personnel response time; Adjusted time for congestion.

[0073] The difference between the remaining time of arrival of the i-th high-risk target and the personnel evacuation time and the preset safety redundancy time is determined as the safety margin of the high-risk target's hazard state. (29) In the formula, Preset safety redundancy time; Indicates the first The time safety margin for a high-risk target to experience secondary combustion or explosion before personnel are evacuated. If This indicates that, after considering safety margin time, the high-risk target is not expected to reach a dangerous state earlier than the time when personnel are evacuated; if This indicates that the high-risk target may enter a state of secondary combustion or explosion hazard before personnel are rescued.

[0074] When multiple high-risk targets exist, the minimum value among the safety margins of the hazard states of each high-risk target is determined as the current safety margin of the hazard state of the high-risk target. (30) when When this occurs, it indicates that at least one high-risk target poses a risk of forming a secondary fire source, causing an explosion, or entering other dangerous states before personnel are freed. The system needs to select subsequent control actions to adjust the longitudinal ventilation speed and the opening and closing status of the lateral smoke exhaust vents in order to extend the remaining time before the danger arrives, reduce the heat flow growth of the high-risk target, and control the risk of secondary disasters. This is the second preset threshold.

[0075] If the dangerous state is not reached within the predicted time domain, the remaining time for the corresponding danger to arrive is recorded as a right censoring value. In this case, instead of arbitrarily assigning a finite value to the remaining time for the danger to arrive, a lower bound for the safety margin of the dangerous state is formed based on the end point of the predicted time domain: (31) In the formula, For the first The candidate action is the first The lower bound of the right-hand deletion safety margin for high-risk targets. Only when this lower bound is not lower than a second preset threshold. When the corresponding state is determined to meet the hard constraint of the dangerous state of the high-risk target, it is determined that the state meets the constraint.

[0076] The predicted time domain length should, in principle, satisfy the following: (32) When the prediction time domain cannot be extended due to limitations in computing resources or the applicability of the model, a conservative judgment is made based on the lower bound of the safety margin, not only because the corresponding state has not reached a dangerous state within the prediction time domain, but also because the corresponding state is deemed safe.

[0077] When the current flue gas layer is stable and has a safety margin Below the first preset threshold Or the safety margin of the current high-risk target's dangerous state. Below the second preset threshold At that time, construct a set of candidate control actions: (33) In the formula, the first Each candidate control action is represented as: (34) In the formula, For the first Candidate longitudinal wind speeds For the first One candidate combination of lateral exhaust port opening and closing.

[0078] For a tunnel with R lateral smoke vents, the candidate lateral smoke vent opening / closing combinations can be expressed as: (35) In the formula, , Indicates the first The side exhaust vent is in the first Activated under candidate control actions. This indicates that the window is closed.

[0079] For each candidate control action, the candidate longitudinal wind velocity and the candidate lateral flue gas outlet state are used as control inputs to the flue gas state prediction model and the surface heat flux prediction model. The corresponding prediction and evaluation relationship is expressed as follows: (36) In the formula, To execute the first Predict the sequence of flue gas layer stability safety margins in the time domain after each candidate control action. For the safety margin sequence of the dangerous state of high-risk targets, For the first A sequence of predicted heat flow on the surface of high-risk targets.

[0080] For the Candidate control actions and prediction times Based on the predicted personnel location, exit status, and path status under this action, the system recalculates the remaining escape time for each trapped person from the predicted time. Simultaneously, based on the predicted heat flow, cumulative heat dose, and hazard criteria of high-risk targets, the first... The remaining time from the predicted time for a high-risk target to reach a dangerous state. .

[0081] The safety margin of the hazard state of the i-th high-risk target at the corresponding predicted time is: (37) The safety margin for the high-risk target's dangerous state at the corresponding predicted time is: (38) For the p-th candidate control action, at each prediction time, the flue gas layer stability safety margin of each control unit is determined according to the aforementioned flue gas layer stability safety margin calculation relationship, based on the flue gas layer height, visibility, and flue gas concentration of each control unit covered by the passable evacuation path. Then, among the control units covered by the passable evacuation path at the corresponding prediction time, the minimum value of the flue gas layer stability safety margin of each control unit is taken as the flue gas layer stability safety margin at that prediction time. .

[0082] A sequence of flue gas layer stability safety margins is formed from the flue gas layer stability safety margins at each prediction time, and the minimum value in the prediction time domain is determined as the [number]th [value]. The worst-case prediction safety margin for flue gas stability for each candidate control action: (39) The secondary disaster risk assessment quantity for high-risk targets is expressed as: (40) In the formula, The most unfavorable predicted flue gas layer stability safety margin for the p-th candidate control action in the prediction time domain; N represents the safety margin of the most unfavorable predicted high-risk target dangerous state for the p-th candidate control action in the prediction time domain; p To predict the number of discrete time steps in the time domain.

[0083] Based on the surface heat flux growth rate of each high-risk target, the roof temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target, the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined: (41) In the formula, The heat flow growth rate of the i-th high-risk target; This is the control unit corresponding to the target; The ceiling temperature gradient for this control unit; This represents the current heat release rate from the heat source. The distance between the fire source and the target; This is the wind speed limiting function calibrated through experiments or simulations.

[0084] The minimum value among the longitudinal wind speed limits corresponding to each high-risk target is determined as the global longitudinal wind speed limit: (42) When the longitudinal wind speed has a non-monotonic relationship with the dangerous state of the high-risk target, the candidate longitudinal wind speeds are evaluated step by step from low to high. The minimum longitudinal wind speed at which the unacceptable dangerous state first appears is used as the judgment boundary. Candidate control actions with a longitudinal wind speed not lower than the minimum longitudinal wind speed are removed from the candidate control action set.

[0085] When the discrete candidate longitudinal wind speeds include 0 m / s, 1 m / s, 2 m / s and 3 m / s, the longitudinal wind speed limit rules can be adopted as shown in Table 2.

[0086] Table 2 Longitudinal Wind Speed ​​Limiting Rules The 3 m / s in Table 2 is only the upper limit of longitudinal wind speed covered by the corresponding scaled-down test and is not a general control limit for actual tunnels. In actual tunnel use, the candidate longitudinal wind speed level should be determined based on the tunnel cross-section, the capacity of the longitudinal ventilation equipment, the fire source status, and the project calibration results.

[0087] From the candidate control actions that meet the global longitudinal wind speed upper limit requirement, select candidate control actions whose worst-case prediction of flue gas layer stability safety margin is not lower than the first preset threshold and whose worst-case prediction of high-risk target dangerous state safety margin is not lower than the second preset threshold, to form a set of feasible control actions: (43) When the set of feasible control actions is not empty, priority is given to candidate control actions that do not reach the dangerous state for each high-risk target in the prediction time domain and whose lower bound of the safety margin of the dangerous state meets the second preset threshold.

[0088] Among candidate control actions with the same right-censoring category, the target control action can be determined based on the overall surface heat flux level of the high-risk target in the predicted time domain, the predicted terminal heat flux or cumulative heat dose, and the change in control action. The corresponding evaluation quantity can be expressed as: (44) (45) (46) in, This is to predict the overall level of surface heat flux of high-risk targets in the time domain; and These are the normalized predicted terminal heat flux and cumulative heat flux, which are then weighted according to their respective weights to form... Without using right censorship, the risk of reaching the reciprocal of the remaining time is not present; Penalty items for controlling changes in actions; , , , and All weights are non-negative. The above evaluation parameters are only used for the optimal selection among candidate control actions that already satisfy the double safety hard constraints and have the same right censoring category.

[0089] When there are no candidate control actions that simultaneously satisfy the first and second preset thresholds, instead of using a weighted approach to cancel out the unsafe state on the personnel escape side and the favorable state on the high-risk target side, the candidate control actions that satisfy the global longitudinal wind speed upper limit are sorted step by step.

[0090] First, compare the degree to which personnel escape safety constraints are met; if the degree to which personnel escape safety constraints are met is the same or the difference is within a preset allowable range, compare the safety margin of the most unfavorable predicted high-risk target dangerous state; then compare the predicted terminal heat flux or cumulative heat flux; subsequently compare the surface heat flux growth rate; finally compare the change in control action. Only when the results of the previous level comparison are the same or the difference is within a preset allowable range will the next level comparison be performed.

[0091] Candidate control actions for which none of the high-risk targets within the prediction time domain have reached a dangerous state and whose lower bound of the dangerous state safety margin meets the second preset threshold are given priority over candidate control actions for which a dangerous state has already been reached within the prediction time domain. Through the above hierarchical sorting, target control actions are determined in order of priority for the degree of satisfaction of personnel escape safety constraints over the degree of satisfaction of secondary disaster risk control constraints of high-risk targets.

[0092] The target control action can be represented as: (47) In the formula, For the target longitudinal wind speed, The target side exhaust port is in open or closed state.

[0093] The system controls the longitudinal ventilation equipment and lateral smoke exhaust outlets to perform target control actions. The longitudinal ventilation equipment adjusts the fan operating frequency, start / stop status, or damper opening according to the target longitudinal wind speed. The lateral smoke exhaust outlet actuators drive the corresponding smoke exhaust outlets to open or close according to the target lateral smoke exhaust outlet opening / closing combination.

[0094] The system collects data on actual longitudinal wind speed, the operating status of longitudinal ventilation equipment, and the position of lateral smoke exhaust vents, and compares the actual execution status with the target control actions. An execution anomaly is determined when the deviation between the actual and target longitudinal wind speed exceeds the preset tolerance, the longitudinal ventilation equipment fails to reach the target operating status, the lateral smoke exhaust vent fails to reach the target opening / closing position within a preset time, the actuator malfunctions, or the corresponding action becomes unexecuted.

[0095] For longitudinal ventilation equipment status and lateral smoke exhaust status that have failed or are currently unexecutable, candidate control actions containing the corresponding failed equipment status or unexecutable status will be removed from the candidate control action set, and the available equipment status and candidate control action set will be updated.

[0096] When the updated set of candidate control actions is empty, enter the preset fail-safe mode: adjust the longitudinal wind speed to the nearest controllable level that the equipment can achieve and does not exceed the global longitudinal wind speed limit, maintain the state of the lateral smoke exhaust outlet that is conducive to smoke exhaust through a passable evacuation path and is still executable, and output manual intervention, fire linkage and emergency evacuation enhancement commands; after the fail-safe action is executed, continue to collect feedback and conduct rolling evaluation.

[0097] Based on the updated fire status data, available equipment status, and candidate control action set, the flue gas state, surface heat flux timing, and dual safety margin corresponding to the candidate control actions are re-predicted, and the target control actions are re-determined.

[0098] Rolling updates can be triggered by any of the following conditions: the preset control period is reached, a significant change in the fire situation occurs, or an execution feedback anomaly occurs. The updated fire situation data can be represented as follows: (48) The rolling closed-loop control process can be represented as: (49) If the updated current flue gas layer stability safety margin is still lower than the first preset threshold, or the updated current high-risk target danger state safety margin is still lower than the second preset threshold, the candidate control action prediction evaluation and target control action determination will continue.

[0099] When the current smoke layer stability safety margin is not lower than the first preset threshold, and the current high-risk target danger status safety margin is not lower than the second preset threshold, the current ventilation and smoke extraction status is maintained and fire scene status data is continuously acquired. During the maintenance of the current ventilation and smoke extraction status, the status of the fire source, the status of high-risk targets, the location of trapped personnel, the status of safety exits, and the status of execution equipment are still updated according to the preset sampling frequency.

[0100] During the rolling control process, the status of personnel evacuation and the risk of secondary disasters at high-risk targets can be continuously assessed. The evacuation is considered complete when the tunnel monitoring system, evacuation guidance system, or escape door monitoring system confirms that all trapped personnel have left the danger zone.

[0101] When direct confirmation cannot be obtained, the personnel evacuation time countdown serves only as an auxiliary judgment, simultaneously assessing whether trapped personnel are still detected within the danger zone, whether the evacuation route status has adversely changed, and whether safety exits remain available. If any of these conditions are not met, the evacuation is not deemed complete solely because the personnel evacuation time countdown has ended.

[0102] When the surface heat flux of a high-risk target decreases or remains stable, the high-risk target does not reach a dangerous state within the prediction time domain and the lower bound of the corresponding dangerous state safety margin meets the second preset threshold, and the current dangerous state safety margin of the high-risk target is not lower than the second preset threshold, it can be determined that the secondary disaster risk of the high-risk target has reached the preset elimination condition.

[0103] Once personnel have been rescued, the risk of secondary disasters from high-risk targets has reached the preset elimination conditions, and the current safety margin of the flue gas layer is not lower than the first preset threshold, and the preset safety maintenance time has been maintained continuously, the longitudinal ventilation equipment and lateral smoke exhaust outlets are gradually restored to the normal smoke exhaust state or the post-accident ventilation state.

[0104] If personnel have been rescued but high-risk targets still pose a risk of secondary disasters, secondary disaster risk control can continue to be implemented for high-risk targets by limiting longitudinal wind speed, adjusting the combination of lateral smoke exhaust port openings, and reducing the heat flow effect in the target area. When the flue gas layer stability safety margin decreases again, the surface heat flow of high-risk targets increases again, or new high-risk targets are identified, the candidate control action prediction and evaluation process is restarted.

[0105] Example 2 This embodiment illustrates the basic experimental data and calibration method of the heat flow time series database. The existing verification data comes from a scaled-down dual-ethanol oil bath test under longitudinal ventilation conditions. The verification scope is scaled-down test verification and mechanism simulation explanation, and the test is not described as verification for actual operational tunnel engineering.

[0106] The experiment was conducted on a 1:3 scaled rectangular cross-section tunnel model. The tunnel model had a cross-sectional dimension of 1m × 2m and a total length of 50m. An axial flow variable frequency fan was installed at one end, and the longitudinal wind speed was adjustable within the range of 0 to 3m / s.

[0107] Square steel oil tanks with sides of 30cm and 40cm and a height of 6cm were used. The combustible material was anhydrous ethanol with a purity of 99.7%. The two oil tanks were arranged along the centerline of the tunnel, with edge spacing of 17.5cm, 35cm, 52.5cm and 70cm respectively.

[0108] The upstream and downstream oil tank dimensions include two configurations: 40cm upstream and 30cm downstream, and 30cm upstream and 40cm downstream. Longitudinal wind velocities were set to 0m / s, 1m / s, 2m / s, and 3m / s, respectively. These two oil tank configurations, four spacing options, and four longitudinal wind speed combinations formed 32 test conditions.

[0109] Heat flux sensors were deployed around the downstream oil pool to collect four-way heat flux. An electronic balance with a range of 35 kg and a resolution of 0.1 g was installed below the upstream oil pool to collect mass loss data at a frequency of 1 Hz. A charge-coupled device (CCD) camera was used to record the flame and ignition process at 25 frames per second. A 50 kW / m² range was used. 2 Heat flux was measured using a circular foil heat flux meter. Temperature was acquired at a frequency of 1 Hz using a K-type armored thermocouple with a diameter of 1 mm.

[0110] The ignition results and ignition times of the 32 groups of tests are shown in Table 3. "Unignited" indicates that no ignition occurred during the corresponding test observation period, and is not arbitrarily assigned the remaining time until the finite danger arrives.

[0111] Table 3 Results of scaled-down dual oil bath test Of the 32 groups tested, 14 groups ignited and 18 groups did not. Statistically, the number of ignition groups at distances of S=17.5cm, 35cm, 52.5cm, and 70cm were 7 / 8, 6 / 8, 1 / 8, and 0 / 8, respectively; 8 / 16 groups ignited when the upstream oil pool side length was 40cm, and 6 / 16 groups ignited when the upstream oil pool side length was 30cm.

[0112] The test results show that, within the test coverage area, the distance between the fire source and the target, the longitudinal wind speed, and the size configuration of the upstream and downstream oil pools all affect the ignition state and ignition time of the downstream oil pool. An increase in longitudinal wind speed does not necessarily lead to a monotonically increasing or decreasing ignition time.

[0113] The mass loss rate, heat release rate of the fire source, four-way heat flow, longitudinal wind speed, oil pool size, oil pool spacing, ignition results and ignition time of each test condition are written into the heat flow time series database as the basic calibration data for the dimensions of fire source size, longitudinal wind speed, distance between fire source and target and scale attribute of high-risk target.

[0114] The impact of target height layers, lateral smoke vent conditions, and tunnel geometry on the surface heat flux of high-risk targets in actual tunnels can be expanded into the heat flux time series database using validated numerical simulation data or project calibration data. Outside the database scope, the closest conservative boundary conditions are used, without unconstrained extrapolation based on the finite wind speed range of the aforementioned scaled-down tests.

[0115] Example 3 This embodiment illustrates the process of calculating dual safety margins and selecting candidate control actions. The ignition time of the high-risk target in this embodiment is derived from a scaled-down dual-oil-pool test, while the smoke layer height, visibility, and smoke concentration are derived from the output of the smoke state prediction model. Therefore, this embodiment is an illustrative example based on measured ignition time and the output of the smoke prediction model, and does not present all data as directly measured data.

[0116] The test condition was selected with an upstream oil tank side length of 40cm, a downstream oil tank side length of 30cm, and a distance of 35cm between the oil tank edges. Candidate longitudinal wind velocities were 0m / s, 1m / s, 2m / s, and 3m / s, with all lateral smoke vents kept closed. The personnel evacuation time was set to 35s, the preset safety redundancy time to 10s, the prediction time domain to 60s, and both the first and second preset thresholds set to 0.

[0117] The safety margin for the hazardous state of high-risk targets is: (50) If a high-risk target does not ignite within the prediction time domain, its safety margin for the hazardous state is 15 seconds.

[0118] The safety margin for flue gas layer stability is: (51) The evaluation results of different candidate control actions are shown in Table 4.

[0119] Table 4. Illustrative Examples of Double Safety Margins Among them, the height ratio of flue gas is Visibility ratio is The concentration ratio is .

[0120] Under the 0m / s candidate control action, the high-risk target did not ignite within the predicted time domain, and the lower bound of the safety margin of the dangerous state met the second preset threshold. However, the smoke layer height and visibility did not meet the requirements for personnel escape, and the smoke layer stability safety margin was negative. Therefore, this candidate control action did not meet the dual safety constraints.

[0121] Under a candidate control action of 1 m / s, the safety margin for the dangerous state of the high-risk target is 10 s, and the safety margin for flue gas layer stability is 0.10. Both safety margins are not lower than the corresponding preset thresholds.

[0122] Under the candidate control actions of 2m / s and 3m / s, the flue gas condition was improved, but the safety margin of the high-risk target was negative, indicating that the high-risk target may enter a dangerous state before the personnel are able to escape.

[0123] Therefore, in this illustrative example, the candidate control action of adjusting the longitudinal wind speed to 1 m / s and keeping the lateral smoke exhaust vent closed is determined as the target control action.

[0124] In another embodiment, a test condition was selected where the side length of the upstream oil pool was 40cm, the side length of the downstream oil pool was 30cm, and the distance between them was 52.5cm. Ignition did not occur at longitudinal wind speeds of 0m / s, 1m / s, and 2m / s, but ignition occurred at a longitudinal wind speed of 3m / s in 323s.

[0125] When the personnel evacuation time is 300s and the preset safety redundancy time is 60s, the safety margin for the hazardous state of the candidate control action at 3m / s is -37s. When 3m / s is taken as the minimum longitudinal wind speed at which an unacceptable hazardous state first occurs, candidate control actions with longitudinal wind speeds not lower than 3m / s are eliminated, and the target control action is determined from the combination of candidate longitudinal wind speeds lower than 3m / s and the opening and closing of the lateral smoke exhaust vents.

[0126] Furthermore, in a close-range scenario where the upstream oil pool has a side length of 30cm, the downstream oil pool has a side length of 40cm, and the distance between them is 17.5cm, the ignition times corresponding to speeds of 1m / s, 2m / s, and 3m / s are 145s, 11s, and 106s, respectively. When the sum of the personnel evacuation time and the preset safety redundancy time exceeds 145s, none of the above three candidate control actions with wind meet the secondary disaster risk control constraints for high-risk targets.

[0127] When a candidate control action at 0 m / s cannot satisfy the personnel escape safety constraints, there are no candidate control actions that simultaneously satisfy both safety constraints. In this case, within the longitudinal wind speed limit range, the degree of satisfaction of personnel escape safety constraints, the safety margin of the high-risk target's dangerous state, the predicted terminal heat flux or cumulative heat flux, the surface heat flux growth rate, and the change in control action are compared in sequence, and the target control action is determined according to the priority of personnel escape safety.

[0128] Example 4 like Figure 2 As shown, this embodiment provides a secondary disaster risk control system for high-risk targets in tunnel fires, including a multi-source data acquisition module, a two-layer spatial modeling module, a high-risk target equivalent module, a smoke layer stability assessment module, a heat flow prediction and risk module, a wind-assisted fire intensity limiting and control decision module, and an execution and feedback module.

[0129] The multi-source data acquisition module is connected to the ceiling temperature sensor, wind speed sensor, heat flow measurement point or heat flow estimation module, video recognition system, fire detection system, traffic monitoring or evacuation guidance system and escape door monitoring system to obtain fire scene status data of tunnel fire.

[0130] The multi-source data acquisition module is also used to perform timestamp alignment, spatial coordinate mapping, outlier removal, missing value compensation and filtering on different data, and send the processed fire status data to the two-layer spatial modeling module, the high-risk target equivalent module, the flue gas layer stability assessment module and the heat flow prediction and risk module.

[0131] The two-layer spatial modeling module is connected to the multi-source data acquisition module to divide the tunnel space longitudinally into multiple control units. Each control unit is further divided vertically into personnel evacuation layer and smoke layer. The module also establishes the relationship between fire source, high-risk target, trapped personnel, safety exit, temperature measurement point, heat flow measurement point, lateral smoke exhaust outlet, and execution equipment and the corresponding control unit.

[0132] The two-layer space modeling module also determines the passable evacuation route based on the location of the trapped personnel, the current available safety exits, and the status of the evacuation route, and sends the control units covered by the passable evacuation route to the flue gas layer stability assessment module and the heat flow prediction and risk module.

[0133] The high-risk target equivalent module is connected to the multi-source data acquisition module and the two-layer spatial modeling module respectively. It is used to determine one or more high-risk targets, the control unit where each high-risk target is located and the direction of heat exposure based on the vehicle identification results, and obtain the corresponding equivalent size, equivalent heat-exposed area, hazard level, hazard parameters and hazard status criteria from the target type-hazard parameter rule table.

[0134] When the identification confidence level is lower than the preset identification confidence level threshold, the high-risk target equivalence module will correct the high-risk target to a vehicle type of the same size level or higher risk level, and send the conservative equivalent status of the high-risk target to the heat flow prediction and risk module.

[0135] The smoke layer stability assessment module is connected to the multi-source data acquisition module, the two-layer spatial modeling module, and the wind-assisted fire intensity limiting and control decision module. Under the current ventilation and smoke extraction conditions, the smoke layer stability assessment module estimates the smoke layer height, visibility, and smoke concentration in the area covered by the passable evacuation path based on the fire scene status data, determines the current smoke layer stability safety margin, and sends this current safety margin to the wind-assisted fire intensity limiting and control decision module.

[0136] After the wind-assisted fire intensity limiting and control decision module constructs a set of candidate control actions, it sends each candidate control action to the flue gas stability assessment module. The flue gas stability assessment module predicts the flue gas state based on each candidate control action, determines the flue gas stability safety margin sequence and the most unfavorable predicted flue gas stability safety margin for the corresponding candidate control action.

[0137] The heat flow prediction and risk module is connected to the multi-source data acquisition module, the two-layer spatial modeling module, the high-risk target equivalence module, and the wind-assisted fire intensity limiting and control decision module. Based on the heat flow time series database and fire scene status data, the heat flow prediction and risk module predicts the surface heat flow time series of each high-risk target, determines the remaining time of danger arrival based on the danger status criteria, determines the personnel evacuation time based on passable evacuation routes, and determines the safety margin of the current high-risk target's danger status.

[0138] For each candidate control action, the heat flow prediction and risk module predicts the heat flow sequence on the surface of the high-risk target. Based on the remaining time for personnel to escape, the remaining time for the danger to arrive, and the preset safety redundancy time, the safety margin sequence of the high-risk target's dangerous state is determined, and the most unfavorable predicted safety margin of the high-risk target's dangerous state for the corresponding candidate control action is determined.

[0139] The wind-assisted fire intensity limiting and control decision-making module is connected to the flue gas stability assessment module, the heat flow prediction and risk module, and the execution and feedback module, respectively. When the current flue gas stability safety margin is lower than the first preset threshold or the current high-risk target danger state safety margin is lower than the second preset threshold, the wind-assisted fire intensity limiting and control decision-making module constructs a set of candidate control actions.

[0140] The wind-assisted fire spread limiting and control decision module determines the upper limit of longitudinal wind speed for each high-risk target based on the surface heat flow growth rate of the high-risk target, the roof temperature gradient of the control unit where the high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the high-risk target, and determines the minimum value among them as the global upper limit of longitudinal wind speed.

[0141] The wind-induced fire spread limiting and control decision module selects candidate control actions from those that meet the global longitudinal wind speed upper limit requirement, and then determines the target control action based on the selection results. If no candidate control action simultaneously meets the dual safety margin constraints, the target control action is determined in a hierarchical order based on the degree of satisfaction of personnel escape safety constraints.

[0142] When the current smoke layer stability safety margin is not lower than the first preset threshold and the current high-risk target danger state safety margin is not lower than the second preset threshold, the wind-assisted fire intensity limiting and control decision module determines to maintain the current ventilation and smoke exhaust state.

[0143] The execution and feedback module is connected to the wind-assisted fire intensity limiting and control decision module, the longitudinal ventilation equipment, the lateral smoke exhaust outlet actuator, and the multi-source data acquisition module, respectively. It is used to send the target longitudinal wind speed command to the longitudinal ventilation equipment and the target lateral smoke exhaust outlet opening and closing command to the lateral smoke exhaust outlet actuator.

[0144] The execution and feedback module collects the actual longitudinal wind speed, the operating status of the longitudinal ventilation equipment, and the position status of the lateral smoke exhaust outlets. It then feeds the collected results back to the multi-source data acquisition module and the wind-assisted fire intensity limiting and control decision module to update the fire scene status data, the status of available equipment, and the set of candidate control actions. Finally, it re-evaluates the candidate control actions and determines the target control actions.

[0145] While maintaining the current ventilation and smoke extraction status, the execution and feedback module controls the multi-source data acquisition module to continuously acquire fire scene status data, and re-triggers candidate control action evaluation when any current safety margin is lower than the corresponding preset threshold, a new high-risk target is identified, or the equipment execution status becomes abnormal.

[0146] Example 5 This embodiment provides an electronic device. The electronic device includes a processor, a memory, a data interface, and a control interface.

[0147] The memory is connected to the processor and is used to store control programs that can be executed by the processor, tunnel geometry and topology data, target type parameter tables, and heat flow timing databases.

[0148] The data interface is used to receive timestamped fire scene status data, trapped personnel data, safety exit data, and evacuation route data. The data interface can be connected to ceiling temperature sensors, wind speed sensors, heat flow measurement points or heat flow estimation modules, video recognition systems, fire detection systems, traffic monitoring or evacuation guidance systems, and escape door monitoring systems.

[0149] The control interface is used to connect the controller of the longitudinal ventilation equipment and the actuator of the lateral smoke exhaust outlet. The controller of the longitudinal ventilation equipment can be a longitudinal fan frequency converter, start / stop controller, or damper controller. The control interface sends the target longitudinal wind speed command to the longitudinal ventilation equipment, sends the lateral smoke exhaust outlet opening / closing command to the actuator of the lateral smoke exhaust outlet, and receives the actual longitudinal wind speed, the operating status of the longitudinal ventilation equipment, and the position status of the lateral smoke exhaust outlet.

[0150] The processor executes the control program in memory, sequentially completing the acquisition and processing of fire scene status data, two-layer spatial modeling, conservative equivalence of high-risk targets, determination of passable evacuation paths, estimation of smoke state, time-series prediction of surface heat flux, calculation of dual safety margins, construction of candidate control actions, longitudinal wind speed limiting, screening of candidate control actions, execution of target control actions, and execution feedback updates.

[0151] The processor can be implemented using an industrial control computer, a programmable logic controller, or an edge computing device. The electronic equipment can be located in the tunnel monitoring center or at an edge control node that communicates with the tunnel monitoring system.

[0152] Example 6 This embodiment provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program causes the processor to perform the aforementioned method for controlling secondary disaster risks of high-risk targets in tunnel fires.

[0153] The computer program includes at least the following instructions: instructions for acquiring and processing fire scene status data; instructions for establishing two-layer space and control units; instructions for determining high-risk targets and dangerous state criteria; instructions for determining passable evacuation routes; instructions for determining the current dual safety margins; instructions for constructing and predicting candidate control actions; instructions for determining the longitudinal wind speed limit and target control actions; and instructions for rolling updates based on execution feedback.

[0154] The computer-readable storage medium can be located within the memory of the aforementioned electronic device, or it can serve as a separate storage medium for installing or updating the control program. The computer-readable storage medium can be a read-only memory, random access memory, flash memory, hard disk, solid-state memory, or optical storage medium.

[0155] Although embodiments of this application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of this application. The scope of this application is defined by the appended claims and their equivalents, all of which should be included within the protection scope of this application.

Claims

1. A method for controlling secondary disaster risks at high-risk targets in tunnel fires, characterized in that, include: Acquire fire scene status data for tunnel fires; Based on the fire situation data, identify one or more high-risk targets and the corresponding danger status criteria for each high-risk target, and determine the passable evacuation routes based on the location of trapped personnel, the current available safety exits, and the status of evacuation routes; Under the current ventilation and smoke extraction conditions, the smoke layer height, visibility, and smoke concentration in the area covered by the passable evacuation route are estimated based on the fire situation data, and the current smoke layer stability safety margin is determined; the surface heat flow sequence of each high-risk target in the prediction time domain is predicted, and the current high-risk target danger state safety margin is determined. When the current flue gas layer stability safety margin is lower than the first preset threshold, or the current high-risk target danger state safety margin is lower than the second preset threshold, a set of candidate control actions is constructed. Based on each candidate control action, predict the smoke layer height, visibility, and smoke concentration in the coverage area of ​​the passable evacuation path within the prediction time domain, and predict the surface heat flow sequence of each high-risk target within the prediction time domain; determine the smoke layer stability safety margin sequence based on the smoke layer height, visibility, and smoke concentration corresponding to each prediction time, and determine the minimum value in the smoke layer stability safety margin sequence as the most unfavorable predicted smoke layer stability safety margin for the corresponding candidate control action; determine the remaining evacuation time of personnel based on the personnel location, exit status, and path status corresponding to each prediction time under the corresponding candidate control action; determine the remaining arrival time of danger for each high-risk target from the corresponding prediction time based on the surface heat flow sequence and corresponding hazard status criteria; determine the hazard status safety margin sequence of high-risk targets based on the remaining arrival time of danger, the remaining evacuation time of personnel, and the preset safety redundancy time, and determine the minimum value in the hazard status safety margin sequence of high-risk targets as the most unfavorable predicted hazard status safety margin for the corresponding candidate control action; Based on the surface heat flow growth rate of each high-risk target, the roof temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target, the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined, and the minimum value among the upper limits of the longitudinal wind speed corresponding to each high-risk target is determined as the global upper limit of the longitudinal wind speed. From the candidate control actions that meet the global longitudinal wind speed upper limit requirements, select the candidate control actions that have a safety margin of flue gas layer stability of the most unfavorable prediction not lower than the first preset threshold and a safety margin of dangerous state of the most unfavorable prediction not lower than the second preset threshold, and determine the target control action based on the selection results; when there are no candidate control actions that simultaneously meet the first preset threshold and the second preset threshold, determine the target control action in the order of priority of meeting personnel escape safety constraints over meeting secondary disaster risk control constraints of high-risk targets. Control the longitudinal ventilation equipment and lateral smoke exhaust outlets to execute target control actions, collect actual longitudinal wind speed, longitudinal ventilation equipment operating status and lateral smoke exhaust outlet positioning status, update fire scene status data, available equipment status and candidate control action set based on the collected results, and re-predict and evaluate candidate control actions and determine target control actions; When the current smoke layer stability safety margin is not lower than the first preset threshold and the current high-risk target danger state safety margin is not lower than the second preset threshold, the current ventilation and smoke exhaust state is maintained and fire status data is continuously acquired.

2. The method for controlling secondary disaster risks of high-risk targets in tunnel fires according to claim 1, characterized in that, Fire scene status data includes roof temperature field, heat flow of the target or adjacent area, longitudinal wind speed, status of lateral smoke exhaust outlets, vehicle identification results, fire source location and fire source heat release rate, location and evacuation status of trapped personnel, status of currently available safety exits, status of evacuation routes, and tunnel geometry and topology data. The fire scene status data is processed by timestamp alignment, spatial coordinate mapping, outlier removal, missing value compensation, and filtering. The tunnel space is divided into multiple control units along the longitudinal direction, and each control unit is further divided into personnel evacuation layer and smoke layer along the height direction. Fire sources, high-risk targets, trapped personnel, safety exits, temperature measuring points, heat flow measuring points, lateral smoke exhaust outlets, and execution equipment are mapped to the corresponding control units according to the longitudinal coordinates, the height layer they are located in, and the boundary of the control unit. For an object located at the boundary of an adjacent control unit, map it to a control unit that is closer to the object; when the spatial influence range of the same object covers multiple control units, establish the association between the object and the corresponding control unit respectively.

3. The method for controlling secondary disaster risks of high-risk targets in tunnel fires according to claim 1, characterized in that, Vehicle identification results include vehicle type, spatial location, operating status, stationary status, and identification confidence level; Based on the vehicle identification results, the target type, spatial location, and identification confidence level of the high-risk target are determined, and the heat exposure direction of the high-risk target is determined based on the spatial relationship between the fire source and the high-risk target. Obtain the equivalent size, equivalent heated area, hazard level, hazard parameters, threshold source and hazard status criterion corresponding to the target type from the target type-hazard parameter rule table, and convert high-risk targets into potential secondary ignition sources or hazard status targets. Hazard parameters include critical surface heat flux, critical cumulative heat dose and critical surface temperature. When the identification confidence level is lower than the preset identification confidence level threshold, the high-risk target is corrected to a vehicle type of the same size level or higher risk level, and the risk parameters and risk status criteria corresponding to the corrected vehicle type are adopted.

4. The method for controlling secondary disaster risks of high-risk targets in tunnel fires according to claim 1, characterized in that, For each control unit covered by the passable evacuation route, the smoke layer height sub-margin is determined based on the normalized deviation of the smoke layer height relative to the minimum safe smoke layer height, the visibility sub-margin is determined based on the normalized deviation of the visibility relative to the minimum safe visibility, and the smoke concentration sub-margin is determined based on the normalized deviation of the smoke concentration relative to the safe threshold of the smoke concentration. The minimum value among the flue gas layer height sub-margin, visibility sub-margin, and flue gas concentration sub-margin is determined as the flue gas layer stability safety margin of the corresponding control unit, and the minimum value among the flue gas layer stability safety margins corresponding to each prediction time and each control unit in the prediction time domain is determined as the most unfavorable predicted flue gas layer stability safety margin. The auxiliary stability ranking index is determined based on the ceiling temperature gradient and longitudinal wind speed disturbance between adjacent control units. The auxiliary stability ranking index is used to rank the candidate control actions that meet the personnel escape safety constraints when determining the target control action.

5. The method for controlling secondary disaster risks of high-risk targets in tunnel fires according to claim 1, characterized in that, The surface heat flow time series is determined based on a pre-established heat flow time series database, which is indexed by the heat release rate of the fire source, the longitudinal wind speed, the distance between the fire source and the high-risk target, the height layer where the high-risk target is located, the status of the lateral smoke exhaust outlet, and the scale attributes of the high-risk target. When the heat release rate of the fire source, longitudinal wind speed, distance between the fire source and the high-risk target, height level of the high-risk target, status of the lateral smoke exhaust outlet, and scale attributes of the current fire condition are within the coverage of the typical fire condition target surface heat flow time series database, the baseline surface heat flow time series is obtained by nearest neighbor condition matching or multilinear interpolation. When any current operating condition parameter exceeds the coverage of the typical operating condition target surface heat flow time series database, the surface heat flow time series corresponding to the conservative boundary condition that is closest to the current fire condition is adopted. Based on the target type and heat exposure direction of high-risk targets, target type correction coefficients and heat exposure direction correction coefficients are determined to correct the baseline surface heat flow time series. Based on the changes in roof temperature, roof temperature gradient, longitudinal wind speed, lateral smoke exhaust port opening and closing, and historical heat flow data of the target surface in the control unit where the corresponding high-risk target is located, residual correction is performed on the corrected surface heat flow time series, and the residual correction amount is continuously updated according to the deviation between the latest measured heat flow and the predicted heat flow.

6. The method for controlling secondary disaster risks of high-risk targets in tunnel fires according to claim 1, characterized in that, Based on the surface heat flux time series and sampling time interval of each high-risk target, the cumulative heat dose of the corresponding high-risk target in the prediction time domain is determined; The hazard condition criteria include one or more of the following: the target surface heat flux reaches or exceeds the critical surface heat flux, the cumulative heat dose reaches or exceeds the critical cumulative heat dose, and the target surface temperature reaches or exceeds the critical surface temperature. The time between the current evaluation time and the earliest predicted time that the hazard condition criteria are met is determined as the hazard arrival time. For each trapped person, paths and safety exits that are blocked by fire sources, have smoke environments that do not meet the passage conditions, or are closed are removed from the tunnel topology map. Based on the path length, the evacuation speed of the person and the congestion status of the path, the evacuation time of the trapped person to each currently available safety exit is determined, and the minimum value among them is determined as the evacuation time of the trapped person. The maximum evacuation time for all trapped personnel is determined by the sum of the personnel reaction time and the congestion correction time. The difference between the remaining time of danger arrival and the time of personnel evacuation and the preset safety redundancy time is determined as the safety margin of the corresponding high-risk target's dangerous state. If the dangerous state is not reached within the prediction time domain, the remaining time for the corresponding danger to arrive is recorded as the right censoring quantity. The lower bound of the safety margin of the dangerous state is determined based on the duration from the current evaluation time to the end of the prediction time domain, the personnel evacuation time, and the preset safety redundancy time. Only when the lower bound of the safety margin of the dangerous state is not lower than the second preset threshold is the corresponding candidate control action determined to meet the hard constraint of the dangerous state of the high-risk target.

7. The method for controlling secondary disaster risks of high-risk targets in tunnel fires according to claim 1, characterized in that, Based on the surface heat flux growth rate of each high-risk target, the ceiling temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target, the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined from the wind speed limiting function calibrated by fire test or fire simulation, and the minimum value of the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined as the global upper limit of the longitudinal wind speed. When the longitudinal wind speed has a non-monotonic relationship with the dangerous state of the high-risk target, the minimum longitudinal wind speed at which the unacceptable dangerous state first appears is used as the boundary to remove candidate control actions from the candidate control action set whose candidate longitudinal wind speed is not lower than the minimum longitudinal wind speed. When the candidate control actions that simultaneously satisfy the first preset threshold and the second preset threshold are not empty, the candidate control actions that do not reach the dangerous state of each high-risk target in the prediction time domain and whose lower limit of the safety margin of the dangerous state satisfies the second preset threshold are selected first. Among the candidate control actions with the same right censoring category, the target control action is determined according to the overall level of surface heat flow of the high-risk target in the prediction time domain, the predicted terminal heat flow or cumulative heat dose, and the change in control action. When there are no candidate control actions that simultaneously meet the first and second preset thresholds, the candidate control actions that meet the global longitudinal wind speed upper limit requirements are sorted in order of the degree of satisfaction of personnel escape safety constraints, the safety margin of the most unfavorable predicted high-risk target dangerous state, the predicted terminal heat flow or cumulative heat dose, the surface heat flow growth rate and the change in control action, and the target control action is determined based on the sorting results. When the execution feedback indicates that the deviation between the actual longitudinal wind speed and the target longitudinal wind speed exceeds the preset tolerance, the longitudinal ventilation equipment has not reached the target operating state, the lateral smoke exhaust port has not reached the target opening and closing position within the preset time, the actuator has malfunctioned, or the corresponding action is not executable, the candidate control action containing the corresponding faulty equipment state or unexecutable state is removed from the candidate control action set, and the target control action is re-determined.

8. A secondary disaster risk control system for high-risk targets in tunnel fires, characterized in that, include: A multi-source data acquisition module is used to acquire fire scene status data in tunnel fires; The two-layer spatial modeling module is used to divide the tunnel space into a personnel evacuation layer and a smoke layer, and further divide it into multiple control units along the longitudinal direction of the tunnel. Based on the location of the trapped personnel, the currently available safety exits, and the status of the evacuation routes, the module determines the passable evacuation routes. The high-risk target equivalent module is used to determine one or more high-risk targets, the control unit where each high-risk target is located, and the danger status criteria corresponding to each high-risk target based on fire scene status data. The smoke layer stability assessment module is used to estimate the smoke layer height, visibility, and smoke concentration in the area covered by the passable evacuation path under the current ventilation and smoke exhaust conditions, based on fire state data. It also determines the current smoke layer stability safety margin based on the margins of the smoke layer height, visibility, and smoke concentration relative to the corresponding safety thresholds. Furthermore, it is used to predict the smoke state for each candidate control action, determine the smoke layer stability safety margin sequence based on the predicted smoke state, and determine the minimum value in the smoke layer stability safety margin sequence as the most unfavorable predicted smoke layer stability safety margin for the corresponding candidate control action. The heat flow prediction and risk module is used to predict the surface heat flow sequence of each high-risk target in the prediction time domain. Based on the surface heat flow sequence and the corresponding hazard status criteria, it determines the remaining time of arrival of the hazard for each high-risk target, determines the personnel evacuation time based on the passable evacuation path, and determines the safety margin of the current high-risk target's hazard status based on the difference between the remaining time of arrival of the hazard and the personnel evacuation time and the preset safety redundancy time. It is also used to predict the surface heat flow sequence of high-risk targets for each candidate control action, determine the safety margin sequence of the high-risk target's hazard status based on the remaining personnel evacuation time, the remaining time of arrival of the hazard, and the preset safety redundancy time, and determine the minimum value in the high-risk target's hazard status safety margin sequence as the most unfavorable predicted high-risk target hazard status safety margin for the corresponding candidate control action. The wind-assisted fire spread limiting and control decision module is used to construct a set of candidate control actions when the current smoke layer stability safety margin is lower than a first preset threshold or the current high-risk target danger state safety margin is lower than a second preset threshold. Each candidate control action includes a candidate longitudinal wind speed and a candidate combination of opening and closing lateral smoke exhaust ports. Based on the surface heat flux growth rate of each high-risk target, the roof temperature gradient of the control unit where the corresponding high-risk target is located, the heat release rate of the fire source, and the distance between the fire source and the corresponding high-risk target, the upper limit of the longitudinal wind speed corresponding to each high-risk target is determined, and the minimum value among the upper limits of the longitudinal wind speed corresponding to each target is determined as the global upper limit of the longitudinal wind speed. The module then selects the target that meets the global upper limit of the longitudinal wind speed. Among the required candidate control actions, those with the most unfavorable predicted flue gas layer stability safety margin not lower than the first preset threshold and the most unfavorable predicted high-risk target danger state safety margin not lower than the second preset threshold are selected, and the target control action is determined based on the selection results; when there are no candidate control actions that simultaneously meet the first and second preset thresholds, the target control action is determined in the order of priority of personnel escape safety constraints over high-risk target secondary disaster risk control constraints; when the current flue gas layer stability safety margin is not lower than the first preset threshold and the current high-risk target danger state safety margin is not lower than the second preset threshold, the current ventilation and smoke exhaust state is maintained. The execution and feedback module is used to control the longitudinal ventilation equipment and lateral smoke exhaust outlets to perform target control actions, collect actual longitudinal wind speed, longitudinal ventilation equipment operating status, and lateral smoke exhaust outlet positioning status, and feed the collected results back to the multi-source data acquisition module and the wind-assisted fire intensity limiting and control decision module to update fire scene status data, available equipment status, and candidate control action set, and to re-predict and evaluate candidate control actions and determine target control actions; it is also used to control the multi-source data acquisition module to continuously acquire fire scene status data while maintaining the current ventilation and smoke exhaust status.

9. An electronic device, characterized in that, Includes processor, memory, data interface, and control interface; The memory is connected to the processor and is used to store control programs that can be executed by the processor, tunnel geometry and topology data, target type parameter tables, and heat flow timing databases; The data interface is used to receive timestamped fire status data, trapped personnel data, safety exit data, and evacuation route data; The control interface is used to connect the controller of the longitudinal ventilation equipment and the actuator of the lateral smoke exhaust outlet, send the target longitudinal wind speed and the opening and closing commands of the lateral smoke exhaust outlet, and receive the execution status feedback; The processor is used to execute a control program to cause the electronic device to perform the secondary disaster risk control method for high-risk targets in tunnel fires as described in any one of claims 1 to 7.

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 method for controlling secondary disaster risks of high-risk targets in tunnel fires as described in any one of claims 1 to 7.

Images