Cable tunnel unmanned aerial vehicle fire extinguishing method and system with emergency intervention and inspection early warning functions

Abstract

The invention discloses a cable tunnel unmanned aerial vehicle fire extinguishing method and system with emergency intervention and inspection early warning functions. The method comprises the steps that environment point cloud data, motion state data of an unmanned aerial vehicle and two-dimensional thermal imaging frame data are collected in real time; constructing and updating a digital twinborn model containing the three-dimensional geometric structure and the thermal field distribution of the cable tunnel; if an abnormal heat source is detected and the comprehensive early warning index exceeds a preset threshold value, an emergency fire behavior is triggered, abnormal heat source analysis is carried out, trajectory correction is carried out, and a fire extinguishing device is guided to carry out fixed-point elimination operation, otherwise, a temperature rise trend is analyzed, and potential hidden danger points are identified to carry out early warning; and filing and storing key data, model snapshots and result reports formed in the whole process for post redisk analysis and model optimization. According to the invention, early warning, accurate positioning and intelligent disposal of fire can be realized, and full-chain closed-loop management from hidden danger identification to post analysis is formed, so that the fire prevention and control capability and the safety operation and maintenance level of the cable tunnel are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of cable tunnel fire prevention and control technology, and relates to a method and system for extinguishing cable tunnel fires using unmanned aerial vehicles (UAVs) that combines emergency intervention and inspection and early warning. Background Technology

[0002] As a crucial underground infrastructure in cities, cable tunnels have always been a focus of industry attention regarding fire safety. Currently, fire prevention and control in cable tunnels primarily relies on fixed sensor networks and track-mounted mobile robots for inspection and early warning. For example, some studies have proposed deploying guide rails and drive devices within the tunnel to mount dry powder fire extinguishers, control boxes, and other equipment on mobile carts for inspection and fire suppression. Other solutions utilize track-mounted robots equipped with cameras and infrared 3D imaging devices to detect the tunnel environment and quickly extinguish small fires. While these methods can achieve fire monitoring and response to some extent, they are limited by track deployment and robot movement, making it difficult to achieve comprehensive coverage and rapid response across the entire tunnel space.

[0003] In recent years, fire source localization methods based on SLAM (Simultaneous Localization and Mapping) technology have gradually emerged. Some studies utilize lidar and SLAM algorithms to construct real-time maps and identify fire sources by capturing images with cameras, converting image coordinates into actual coordinates to achieve spatial localization of the fire source. Other approaches propose calibrating visible light cameras, infrared thermal imagers, and depth cameras, fusing temperature field and depth information to achieve accurate identification and spatial localization of mobile heat sources. While these technologies have made some progress in fire source localization, most remain limited to ground platforms and do not fully utilize the maneuverability of unmanned aerial vehicles (UAVs) to cope with complex tunnel environments.

[0004] Regarding the collaborative operation of drones and firefighting robots, existing research has proposed establishing a three-dimensional spatial geometric positioning model to acquire the three-dimensional coordinates of the drones and robots in real time, and to perform online compensation for the water cannon spray angle to improve accuracy. However, these solutions are mainly aimed at open outdoor scenarios and do not consider the special characteristics of enclosed spaces such as cable tunnels. Furthermore, existing technologies generally suffer from a disconnect between emergency response and routine inspections, lacking comprehensive closed-loop management. In particular, there is still significant room for improvement in utilizing digital twin technology for time-series temperature field analysis and achieving seamless integration from routine inspections to emergency response.

[0005] Existing cable tunnel fire monitoring systems suffer from problems such as untimely early warnings and low accuracy in locating fire sources. Traditional fixed sensor networks have limited coverage, making it difficult to effectively track dynamic heat sources. Meanwhile, in tunnel environments where GPS signals are limited, unmanned aerial vehicle (UAV) platforms face challenges in navigation and positioning accuracy, affecting the accurate location of fire sources and timely intervention. These problems often result in missed golden hours for fire response, increasing the risk of fire spread.

[0006] Furthermore, traditional cable tunnel fire prevention and control systems suffer from a disconnect between emergency response and post-incident analysis, lacking a comprehensive closed-loop management mechanism. Existing technologies struggle to achieve seamless integration from hazard warning and real-time monitoring to emergency response, resulting in low overall system response efficiency and an inability to adapt to complex and ever-changing fire situations. Moreover, traditional methods are deficient in the fusion and utilization of multi-source heterogeneous data, hindering comprehensive perception and accurate assessment of the fire situation.

[0007] Finally, existing technologies still need improvement in autonomous navigation, precision firefighting, and intelligent decision-making for drones in the complex environment of cable tunnels. In particular, key technologies such as stable flight control of drones, dynamic path planning, and precise delivery of firefighting devices have not yet been effectively resolved in narrow, high-temperature, and smoke-filled tunnel environments. These issues limit the widespread application and effectiveness of drone firefighting systems in critical infrastructure such as cable tunnels. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a method and system for extinguishing fires using unmanned aerial vehicles (UAVs) in cable tunnels, which combines emergency intervention and early warning for inspection. This system enables early warning, precise location, and intelligent handling of fires, and forms a closed-loop management chain from hazard identification to post-incident analysis, thereby significantly improving the fire prevention and control capabilities and safe operation and maintenance level of cable tunnels.

[0009] The present invention adopts the following technical solution.

[0010] The first aspect of this invention proposes a method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs), which combines emergency intervention and inspection / early warning, comprising: S1: The drone enters the cable tunnel and collects environmental point cloud data, drone motion status data and two-dimensional thermal imaging frame data in real time through airborne sensors; S2: Construct and update a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel based on real-time collected data; S3: Based on the digital twin model, locate the abnormal heat source based on the thermal anomaly function. After locating the abnormal heat source, calculate the comprehensive early warning index. If the comprehensive early warning index exceeds the preset threshold, trigger an emergency fire and proceed to S4; otherwise, proceed to S5. S4: Based on the digital twin model, abnormal heat source analysis is performed and trajectory correction is carried out to guide the fire extinguishing device to carry out targeted elimination operations; S5: By comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, analyze the temperature rise trend and identify potential hidden dangers for early warning; S6: Archives key data, model snapshots, and results reports generated throughout the process for post-event review and model optimization.

[0011] Preferably, in S2, a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel is constructed and updated based on the real-time collected data, specifically including: S201: Perform geometric and pose calculations on the real-time collected environmental point cloud data and the motion state data of the UAV to generate a three-dimensional geometric point cloud map of the cable tunnel and the real-time pose of the UAV in the map. S202: Process the real-time acquired two-dimensional thermal imaging frame data, assign a corresponding temperature value to each pixel, and form a two-dimensional temperature matrix. S203: Based on the extrinsic calibration matrix between sensors and the real-time pose of the UAV in the map, calculate the projection relationship of each frame of thermal imaging in three-dimensional space. S204: Project the two-dimensional temperature matrix onto the three-dimensional geometric point cloud map according to the projection relationship, and assign a corresponding temperature attribute value to each three-dimensional point in the three-dimensional geometric point cloud map to form a digital twin model of the current moment containing the three-dimensional geometric structure and thermal field distribution of the cable tunnel.

[0012] Preferably, in S3, the location of the abnormal heat source is located based on the thermal anomaly function, as follows:

[0013]

[0014] in, Location of abnormal heat source; This is a function of thermal anomaly. The set of all points in the current digital twin model; For the temperature field in the digital twin model, and These represent the mean and standard deviation of historical temperature data, respectively.

[0015] Preferably, in S3, the comprehensive early warning index is:

[0016] in, , , These are the weighting coefficients; For temperature field ( and insulation degradation status The fire probability model; F is a logical variable representing a fire event, with a value of 1 indicating that a fire has occurred and a value of 0 indicating that a fire has not occurred; This represents the thermal anomaly function value corresponding to the location of the abnormal heat source. This is a function of thermal anomaly. This represents a volume element.

[0017] Preferably, in S4, abnormal heat source analysis is performed based on a digital twin model, and trajectory correction is applied to guide the fire extinguishing device to carry out targeted elimination operations, specifically including: S401: Upon receiving an emergency fire trigger signal, retrieve the latest digital twin model of the moment the emergency fire was triggered, and based on the comprehensive early warning index... Filter the point cloud in the model to select all comprehensive early warning indices. Points exceeding a preset threshold form an abnormal heat source cloud cluster; S402: Calculate the geometric center coordinates of the abnormal heat source point cloud cluster in three-dimensional space as the target coordinates, and obtain the real-time pose information of the UAV and the relative attitude of the fire extinguishing device. S403: The ballistic correction model calculates the projection direction compensation required to hit the target based on the target coordinates, the real-time pose of the UAV, the relative attitude of the fire extinguishing device, and the environmental parameters obtained by the airborne sensors. This includes the pitch angle correction α and the azimuth angle correction β. S404: Correct the target direction according to the projection direction compensation amount, and adjust the projection axis of the drone and the fire extinguishing device to the corrected target direction; S405: The drone performs the deployment of the airborne fire extinguishing device. After deployment, it continuously monitors the temperature changes in the target area to determine whether the fire has been effectively suppressed.

[0018] Preferably, in S5, by comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, the temperature rise trend is analyzed and potential hidden dangers are identified for early warning, specifically including: S501: The UAV maintains a stable flight state according to the preset trajectory, performs daily inspection tasks, and obtains the digital twin model Model-t0 of the current time and the digital twin model Model-t1 of the previous inspection cycle; S502: Employs a point cloud registration algorithm to spatially align Model-t0 and Model-t1; S503: Perform a difference operation on the aligned Model-t0 and Model-t1 to obtain the temperature change at each location point in the tunnel; S504: Determine whether the temperature change exceeds the sensitivity threshold of the temperature rise trend. If so, the corresponding location point is taken as a potential hidden danger point, and a digital twin model of an earlier inspection cycle is obtained to construct a temperature change time series of the potential hidden danger point. S505: Perform fitting analysis on the temperature change time series, calculate the future temperature rise rate, and predict the time required to reach the alarm temperature threshold based on the rate. S506: Graded early warning based on prediction results: If the predicted time is lower than the set value, a high-priority early warning is generated; otherwise, a general attention notice is generated, and the monitoring frequency of the area corresponding to the potential hazard point is automatically increased in future inspections.

[0019] Preferably, the UAV adopts the following optimized control strategy: A joint dynamic model based on UAV status and cable tunnel thermodynamics And the definition of Hamiltonian function for the optimal control problem of unmanned aerial vehicles. According to the Hamiltonian function To obtain the optimal control input for achieving coordinated optimization of state trajectory and control input. The details are as follows:

[0020]

[0021] in, Let be the system state vector. To control the input vector, Let t be the accompanying variable, and t be time. This represents the immediate cost obtained based on the optimal control problem of unmanned aerial vehicles (UAVs). The weighted quadratic norm representing the state deviation. This represents the energy consumption or risk weight of the control input.

[0022] Preferably, a joint kinetic model of the UAV status and cable tunnel thermodynamics. Specifically as follows:

[0023] in, For fire combustion dynamics response; The thermodynamic enthalpy per unit volume or per unit mass of the composite medium in the cable tunnel; This represents the current temperature field inside the tunnel. For system enthalpy; For insulation degradation parameters; The ambient reference temperature This is the heat conduction term for the temperature field; It is a function relating resistance to temperature; Let be the mass flow rate of the extinguishing agent at time t. Δ is the enthalpy change of the reaction between the extinguishing agent and combustion; Q is the heat release rate of the fire source. Let be the weight coefficient of the i-th sub-control module; Let i be the state variable of the i-th subsystem in the joint dynamics model. Let i be the dynamic response function of the i-th subsystem; These are physical and control parameters.

[0024] A second aspect of this invention proposes a cable tunnel unmanned aerial vehicle (UAV) firefighting system that combines emergency intervention and inspection / early warning, comprising: The data acquisition module is used for the UAV to enter the cable tunnel and collect environmental point cloud data, UAV motion status data and two-dimensional thermal imaging frame data in real time through airborne sensors; The digital twin module is used to build and update a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel based on real-time collected data. The anomaly detection module is used to locate the location of abnormal heat sources based on the thermal anomaly function on the basis of the digital twin model. After locating the abnormal heat source, it calculates the comprehensive early warning index. If the comprehensive early warning index exceeds the preset threshold, it triggers an emergency fire and enters the emergency intervention module; otherwise, it inspects the early warning module. The emergency intervention module is used to analyze abnormal heat sources and correct trajectories based on a digital twin model, guiding fire extinguishing devices to carry out targeted elimination operations. The inspection and early warning module is used to analyze the temperature rise trend and identify potential hidden dangers by comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, so as to issue early warnings. The review and optimization module is used to archive and store key data, model snapshots, and result reports generated throughout the process, for post-event review analysis and model optimization.

[0025] A third aspect of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to perform operations according to the instructions to execute the steps of the method.

[0026] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0027] Compared with the prior art, the beneficial effects of the present invention include at least the following: This invention constructs and updates a digital twin model containing the three-dimensional geometric structure and thermal field distribution of cable tunnels based on real-time collected data. This spatiotemporal digital twin model breaks through the limitations of traditional fixed sensor networks, realizing comprehensive perception and dynamic monitoring of the cable tunnel environment. By fusing multi-source sensor data such as lidar, inertial measurement units, and infrared thermal imagers, a high-precision three-dimensional geometric environment model and spatiotemporal dynamic projection of the temperature field are constructed, providing a solid data foundation for accurate fire source location and fire trend analysis. This not only significantly improves the timeliness of fire early warning but also greatly enhances the accuracy of fire source location, effectively solving the problems of untimely early warning and inaccurate location in existing technologies.

[0028] This invention proposes a fire source location algorithm based on thermal anomalies. It integrates local anomalies, global temperature rise rate, and fire probability models into a unified comprehensive early warning index. Fire situation analysis is performed based on this comprehensive early warning index, which solves the problem of unreliability of single threshold judgment, realizes multi-dimensional comprehensive decision-making, and provides accurate target information for subsequent fire fighting operations.

[0029] This invention achieves closed-loop management across the entire chain, from hazard warning and precise intervention to post-event analysis, through a digital twin engine, autonomous navigation, and fire suppression execution. This not only improves the fire prevention and control capabilities of cable tunnels but also significantly enhances the efficiency and reliability of safe operation and maintenance, realizing a comprehensive improvement from hazard warning and precise intervention to post-event analysis.

[0030] The dual-modal workflow and optimized control strategy proposed in this invention effectively bridge the technical gap between emergency response and post-event analysis, significantly improving the overall response efficiency and intelligence level of the system. This invention integrates fire combustion dynamics, heat conduction equations, extinguishing agent effects, and UAV control inputs into a unified dual-modal coupled dynamics framework, establishing a joint dynamic model of UAV status and cable tunnel thermodynamics. This lays the foundation for achieving coordinated optimal control of UAV trajectory and firefighting mission, significantly improving the accuracy of emergency intervention and system efficiency.

[0031] This invention achieves integrated decision-making for UAV inspection and intervention based on spatiotemporal digital twins. The flexibility and maneuverability of the UAV platform and the real-time perception capabilities of digital twin technology offer significant advantages in coverage and dynamic response compared to solutions using fixed sensor networks. Compared to solutions primarily based on track-based mobile robots, this invention, by introducing autonomous navigation in three-dimensional space and precise fire suppression control, demonstrates superior performance in system response speed and fire source handling efficiency. Particularly in tunnel environments where GPS signals are obstructed, high-precision three-dimensional positioning and map construction are achieved by fusing laser point cloud and inertial navigation data, overcoming the challenge of autonomous navigation for UAVs in complex cable tunnel environments and solving the problem of insufficient navigation and positioning accuracy in complex environments with existing technologies.

[0032] This invention can be widely applied to fire prevention and control of critical infrastructure such as cable tunnels and underground utility tunnels. It can significantly improve the timeliness of fire early warning, increase the accuracy of fire source location, and optimize the efficiency of emergency response. Given its excellent performance in complex environments and comprehensive closed-loop management capabilities, this invention has good application prospects in the safety management of smart city infrastructure. Attached Figure Description

[0033] Figure 1 Comparison of 3D localization errors of abnormal heat sources under different localization algorithms; Figure 2 This is a performance comparison chart of the UAV inspection-intervention integrated decision-making method, the GPS method, and the multi-sensor fusion method of the present invention; Figure 3 This is a flowchart of the cable tunnel fire monitoring and emergency response based on a digital twin model, as described in this invention. Figure 4 This is a flowchart of the construction process of a three-dimensional temperature point cloud for cable tunnels based on multi-source sensor fusion, as described in this invention. Figure 5 This is a schematic diagram of the intelligent fire suppression deployment process based on drones according to the present invention; Figure 6 This invention provides a flowchart for analyzing temperature rise trends and identifying potential hazard points using a drone inspection and early warning mode based on real-time point cloud perception. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this invention are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0035] Embodiment 1 of this invention provides a method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs), combining emergency intervention and early warning systems. This addresses problems in existing technologies such as untimely fire warnings, low accuracy in fire source location, and the disconnect between emergency response and post-incident analysis. The method primarily involves a digital twin engine, autonomous navigation, and fire extinguishing execution. The digital twin engine constructs and updates a three-dimensional geometric environment model and a dynamic spatiotemporal projection of the temperature field of the cable tunnel, providing data support for fire source location and fire trend analysis. Autonomous navigation, based on the digital twin model, integrates laser point cloud and inertial navigation data to achieve high-precision three-dimensional positioning and map construction in environments where GPS signals are blocked. Fire extinguishing execution, based on the three-dimensional fire source location results, invokes a ballistic compensation control model to precisely adjust the firing parameters of the fire extinguishing equipment. Specifically, as shown... Figure 3 As shown, the method includes: S1: The drone enters the cable tunnel and collects environmental point cloud data, drone motion status data and two-dimensional thermal imaging frame data in real time through airborne sensors; More preferably, in this embodiment of the invention, the construction of the digital twin model is the core of the system.

[0036] Option 1) The digital twin model is constructed based on the tunnel spatial coordinate system. Combined with the cable temperature rise model, the temperature field is superimposed onto the geometric model to form a time-series digital twin model. Details are as follows: First, define the tunnel spatial coordinate system. ,in The axis runs along the length of the tunnel. and The axes represent the width and height, respectively. The tunnel geometry model can be represented as:

[0037] in, The length of the tunnel. For width, This is a function of the height of the arched top.

[0038] Combining the cable temperature rise model, the temperature field Superimposed on the geometric model, forming a time-series digital twin model:

[0039] Option 2) In this embodiment of the invention, the coordinate system definition method for constructing the digital twin model can adopt an alternative scheme.

[0040] Specifically, the Cartesian coordinate system can be... Convert to cylindrical coordinate system ,in Indicates radial distance. Indicates the azimuth angle.

[0041] The corresponding geometric model can be reconstructed as ,in Describes the shape function of the tunnel cross-section.

[0042] This alternative approach, by altering the spatial parameterization method, still maintains the three-dimensional temperature field. It has precise projection capabilities and is better suited to the modeling needs of circular cross-section tunnels.

[0043] More preferably, in this embodiment of the invention, a multi-source data fusion strategy is also proposed: The extended Kalman filter (EKF) algorithm can be used to achieve multi-source sensor data fusion.

[0044] Define the state vector of the unmanned aerial vehicle system ,in , , These are the quaternions for position, velocity, and attitude, respectively. and For the bias of the accelerometer and gyroscope.

[0045] The system state prediction equation is:

[0046] in, For the previous moment State estimates, For control input, it refers to the control commands or motion status information of the UAV at time k; For a moment State values, corresponding to the system state vector ; The observation update equations are:

[0047] in, and These are the nonlinear state transition function and the observation function, respectively. and For process noise and observation noise, Let be the state covariance matrix. For Kalman gain, Let k be the observation at time k. To observe the noise covariance matrix, Let h(⋅) be the Jacobian matrix of h(⋅) in the current state.

[0048] The filtering algorithm outputs the high-precision position p, velocity v, and attitude quaternion q of the UAV in the tunnel coordinate system. At the same time, it estimates and corrects the bias of the accelerometer and gyroscope. The results can be used as input data for subsequent digital twin model construction (S201-S204) and abnormal heat source analysis (S402-S403).

[0049] Specifically, the Extended Kalman Filter (EKF) can be replaced by the Unscented Kalman Filter (UKF), which approximates a nonlinear distribution through Sigma point sampling, and the state prediction equation can be rewritten as:

[0050] in For the Sigma point set, Here, n represents the corresponding weight coefficients, and n is the dimension of the system state vector. This alternative avoids the calculation of the Jacobian matrix and maintains positioning accuracy even in strongly nonlinear scenarios.

[0051] Another alternative is to use particle filtering (PF), which achieves posterior probability estimation through importance sampling and is suitable for non-Gaussian noise environments.

[0052] S2: Construct and update a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel based on real-time collected data; More preferably, by fusing data from lidar, inertial measurement unit, and infrared thermal imager after the drone enters the cable tunnel, a digital twin model containing three-dimensional geometry and thermal field distribution is constructed and updated in real time, such as... Figure 4 As shown, it specifically includes: S201: Perform geometric and pose calculations on the real-time collected environmental point cloud data and the motion state data of the UAV to generate a three-dimensional geometric point cloud map of the cable tunnel and the real-time pose of the UAV in the map. Environmental point cloud data and UAV motion state data (including angular velocity and acceleration) are simultaneously acquired by LiDAR and IMU. The SLAM algorithm is used to jointly process these sensor data to generate a 3D point cloud map and the UAV's six degrees of freedom (6-DoF) position data.

[0053] S202: Process the real-time acquired two-dimensional thermal imaging frame data, assign a corresponding temperature value to each pixel, and form a two-dimensional temperature matrix. S203: Based on the extrinsic calibration matrix between sensors and the real-time pose of the UAV in the map, calculate the projection relationship of each frame of thermal imaging in three-dimensional space. S204: Project the two-dimensional temperature matrix onto the three-dimensional geometric point cloud map according to the projection relationship, and assign a corresponding temperature attribute value to each three-dimensional point in the three-dimensional geometric point cloud map to form a digital twin model of the current moment containing the three-dimensional geometric structure and thermal field distribution of the cable tunnel.

[0054] Based on the external parameter relationships between sensors and the UAV's position data, the system calculates the relationship between each 3D point and infrared thermal imaging data, and maps the temperature information to the corresponding point cloud. Combined with the temperature data, the system can analyze the temperature distribution in the environment, thereby better assisting in the identification and classification of obstacles.

[0055] Through this series of data fusion processes, this method can efficiently and accurately identify obstacles in the environment and provide reliable support for subsequent path planning and flight safety.

[0056] S3: Locate the location of abnormal heat sources based on the thermal anomaly function, and calculate the comprehensive early warning index after locating the abnormal heat sources. ,like If the preset threshold is exceeded, an emergency fire is triggered and the system enters S4 (emergency intervention mode); otherwise, it enters S5 (inspection and early warning mode). More preferably, during operation, the system compares and judges the digital twin model with the monitored temperature field data to determine whether there is an abnormal heat source exceeding the preset emergency threshold. When an abnormal heat source is detected and the comprehensive warning index exceeds the preset threshold, the system will enter the emergency intervention mode, implement three-dimensional positioning and call the ballistic correction model to accurately guide the fire extinguishing device to carry out targeted elimination operations. If no emergency fire is found, the system will enter the inspection and early warning mode, identify potential temperature rise trends and thermal anomaly evolution by comparing and analyzing the real-time model with the historical model, and generate inspection reports and early warning information.

[0057] The algorithm for locating abnormal heat sources is as follows: Option 1) Based on the fire formation mechanism, this embodiment of the invention designs a fire source location method based on thermal anomalies. By comparing the mean and standard deviation of real-time temperature with historical temperature data, the fire source location is accurately determined. The thermal anomaly function is defined as follows:

[0058] in, and These represent the mean and standard deviation of historical temperature data, respectively. The estimated location of the fire source is:

[0059] in, The set of all points in the current digital twin model; This refers to the temperature field in a digital twin model.

[0060] Option 2) The decision function of the abnormal heat source localization algorithm can be modified. Specifically, the standardized thermal anomaly function... This can be replaced by a function that represents the ratio of absolute temperature difference to historical peak value:

[0061] in This represents the historical extreme temperature value at the location.

[0062] Option 3) Another alternative is to use multivariate anomaly detection based on the covariance matrix, constructing a Mahalanobis distance function:

[0063] in Let be the spatiotemporal covariance matrix of the temperature field. , That is , .

[0064] The location of the anomalous heat source is estimated to be the point that maximizes the Mahalanobis distance, i.e.:

[0065] This approach improves the robustness of positioning in complex heat conduction scenarios by considering spatial correlation characteristics.

[0066] The early warning decision-making model is as follows: This invention constructs a comprehensive early warning index:

[0067] in, , , These are the weighting coefficients. For temperature field and insulation degradation status The fire probability model has a logical variable F representing the fire occurrence event, where a value of 1 indicates that a fire has occurred and a value of 0 indicates that a fire has not occurred. This represents the thermal anomaly function value corresponding to the location of the abnormal heat source; This represents the cable tunnel region for which volume integration is performed (i.e., the geometric volume covered by the current digital twin model). It is a standard mathematical symbol representing the volume element of a volume integral. This invention integrates local anomalies, global temperature rise rate, and fire probability models into a unified early warning indicator, solving the problem of unreliable single threshold judgment and realizing multi-dimensional comprehensive decision-making.

[0068] Early warning decision model When the preset threshold is exceeded, the system triggers an early warning and initiates an emergency response process.

[0069] S4: Based on the digital twin model, abnormal heat source analysis is performed and trajectory correction is carried out to guide the fire extinguishing device to carry out targeted elimination operations; More preferably, abnormal heat sources are analyzed and ballistic corrections are performed based on a digital twin model to guide the fire extinguishing device to carry out targeted elimination operations, such as... Figure 5 As shown, it specifically includes: S401: Upon receiving an emergency fire trigger signal, retrieve the latest digital twin model of the moment the emergency fire was triggered, and based on the comprehensive early warning index... The point cloud in the model is filtered to identify all high-temperature points (comprehensive early warning index). Points exceeding the alarm temperature threshold form an abnormal heat source cloud cluster; S402: The system applies centroid calculation or RANSAC algorithm to the abnormal heat source point cloud cluster to accurately calculate its geometric center coordinates in three-dimensional space, and obtain the real-time pose information of the UAV and the relative attitude of the fire extinguishing device (including the initial velocity of the fire extinguishing agent). S403: The system takes the target coordinates, UAV attitude, initial velocity of the fire extinguishing agent and environmental parameters (such as airflow velocity in the tunnel) obtained by the airborne sensors as input data, loads them into the ballistic correction model for calculation, and obtains the projection direction compensation required to hit the target based on the model solution results, including pitch angle correction α and azimuth angle correction β. S404: Adjusting the projection axis of the drone and the fire extinguishing device to the corrected direction through attitude control; S405: The drone performs the deployment of extinguishing agents from its onboard fire suppression system, such as dry powder or high-pressure water mist. After deployment, the system continuously monitors temperature changes in the target area using an infrared thermal imager to determine whether the fire has been effectively contained. Data is recorded and reported upon mission completion.

[0070] This method integrates point cloud processing, attitude control, and ballistic correction calculations to enable UAVs to accurately identify fires and perform efficient firefighting operations in complex environments. It has the advantages of fast response speed, high firefighting accuracy, and strong environmental adaptability.

[0071] S5: By comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, analyze the temperature rise trend and identify potential hidden dangers for early warning; More preferably, by comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, the temperature rise trend is analyzed and potential hidden dangers are identified for early warning, such as... Figure 6 As shown, it specifically includes: S501: The UAV maintains a stable flight state according to the preset trajectory, performs daily inspection tasks, and obtains the digital twin model Model-t0 of the current time and the digital twin model Model-t1 of the previous inspection cycle; S502: Employs a point cloud registration algorithm to spatially align Model-t0 and Model-t1; S503: Perform a difference operation on the aligned Model-t0 and Model-t1 to obtain the temperature change at each location point in the tunnel; S504: Determine whether the temperature change exceeds the sensitivity threshold of the temperature rise trend. If so, the corresponding location point is taken as a potential hidden danger point, and a digital twin model of an earlier inspection cycle is obtained to construct a temperature change time series of the potential hidden danger point. S505: Perform fitting analysis on the temperature change time series, calculate the future temperature rise rate, and predict the time required to reach the alarm temperature threshold based on the rate. S506: Graded early warning based on prediction results: If the predicted time is lower than the set value, a high-priority early warning is generated; otherwise, a general attention notice is generated, and the monitoring frequency of the area corresponding to the potential hazard point is automatically increased in future inspections.

[0072] More preferably, the present invention also proposes a dual-modal system workflow: The dual-modal (co-optimization of UAV state trajectory and control input) workflow of this invention integrates fire combustion dynamics, heat conduction equations, extinguishing agent effects, and UAV control input into a unified dual-modal coupled dynamics framework. It establishes a joint dynamic model of UAV state and cable tunnel thermodynamics, which can be described by the following dynamic model, including insulation degradation state. Temperature field evolution Combustion reaction and multi-sensor fusion And many other aspects:

[0073] in, Let be the system state vector. To control the input vector, Describe the dynamic response of fire combustion. These are physical and control parameters; The sensitivity coefficient of insulation degradation to fire combustion rate; The temperature difference coupling coefficient, The density of the medium; Specific heat capacity of the composite medium inside the cable tunnel; Thermal conductivity of the composite medium inside the cable tunnel; Current intensity; It is a function of resistance and temperature. The convective heat transfer coefficient, The efficiency coefficient of fire extinguishing agent utilization. This is the fire extinguishing loss coefficient. is the control weight coefficient of the i-th subsystem in the joint dynamics model, and its specific value can be determined through optimal control theory or engineering debugging. The thermodynamic enthalpy per unit volume or per unit mass of the composite medium in the cable tunnel; For system enthalpy, These are insulation degradation parameters. This represents the current temperature field inside the tunnel. The ambient reference temperature For the heat conduction term of the temperature field, This is a function relating resistance and temperature. The mass flow rate of the extinguishing agent at time t. Let ΔQ represent the enthalpy change of the reaction between the extinguishing agent and combustion, and Q represent the heat release rate from the fire source. Let be the weight coefficient of the i-th sub-control module. Let i be the state variable of the i-th subsystem. Let be the dynamic response function of the i-th subsystem.

[0074] By integrating fire combustion dynamics, heat conduction equations, extinguishing agent effects, and UAV control inputs into a unified dual-modal coupled dynamics framework, a joint dynamics model of UAV status and cable tunnel thermodynamics was established. This lays the foundation for achieving coordinated optimal control of UAV trajectory and firefighting mission, significantly improving the accuracy and system efficiency of emergency intervention.

[0075] More preferably, the present invention also proposes an optimized control strategy: The UAV optimized control strategy defines an objective function that minimizes state deviation and control input energy consumption, and achieves co-optimization of state trajectory and control input through a Hamiltonian function. Specifically, The optimization control objective (UAV optimal control problem) of this invention is defined as follows:

[0076] in, The weighted quadratic norm representing the state deviation. This represents the energy consumption or risk weights used to control the input. Penalty for terminal status. At the end of the control interval The system state refers to the final state of the drone after it has performed an inspection / firefighting mission.

[0077] Based on the above dynamic model and optimal control problem, a Hamiltonian function is defined to achieve the joint optimization of state trajectory and control input:

[0078] in, As an accompanying variable, For immediate cost. The Hamiltonian function formula... , That is, the dynamic model mentioned above Optimize control objectives .

[0079] Hamiltonian function The output is used to solve for the optimal control law. According to Pontryagin's maximum principle, the optimal control input... The following conditions must be met:

[0080] S6: Archives key data, model snapshots, and results reports generated throughout the process for post-event review and model optimization.

[0081] More preferably, key data, model snapshots, and result reports generated throughout the process are archived and stored for post-process review and model optimization.

[0082] Through the above technical solution, the present invention realizes a closed-loop management system covering the entire chain from hazard warning and precise intervention to post-event analysis.

[0083] The introduction of the digital twin engine solved the problem of untimely fire warnings. The multi-source data fusion strategy and abnormal heat source location algorithm improved the accuracy of fire source location. The system's dual-modal workflow and optimized control strategy bridged the technical gap between emergency response and post-event analysis.

[0084] These innovations have effectively improved the fire prevention and control capabilities and safe operation and maintenance level of cable tunnels.

[0085] Specifically, in this embodiment of the invention, a simulation environment based on the MATLAB / Simulink platform was constructed to verify the performance of the proposed cable tunnel UAV fire extinguishing system.

[0086] The emulation hardware uses an Intel Xeon Gold 6248 processor (2.5 GHz), 256 GB DDR4 RAM, and 2 TB NVMe SSD storage.

[0087] The simulation model includes core components such as a digital twin engine, an autonomous navigation module, a multi-source data fusion module, and an intelligent decision-making algorithm module, to realistically reflect the operating characteristics and control behavior of high-voltage cable tunnels.

[0088] The model parameters were set with reference to the technical specifications and operational data of high-voltage cable tunnels to ensure that the simulation results have practical significance.

[0089] Figure 1 This paper presents a comparison of the 3D localization errors of different positioning algorithms for abnormal heat sources. The error range of the digital twin method is 7.43 cm to 9.12 cm, with a median of 8.61 cm; the error range of the geometric model localization method is 26.84 cm to 30.21 cm, with a median of 28.53 cm; and the error range of the 2D thermal imaging localization method is 42.15 cm to 47.89 cm, with a median of 45.98 cm. The comparison shows that the digital twin method significantly outperforms the other two methods in 3D localization accuracy, reducing the localization error by approximately 72.35% and 69.79% compared to the 2D thermal imaging localization method and the geometric model localization method, respectively.

[0090] Figure 2 This demonstrates that the spatiotemporal digital twin integrated decision-making method proposed in this invention significantly outperforms the comparative methods in all three key performance indicators. In terms of detection accuracy, the proposed method achieves 99.5%, a significant advantage over the 94.1% of the GPS method and the 97.3% of multi-sensor fusion. Energy efficiency is 98.1%, exceeding the 84.8% of the GPS method and the 91.9% of the multi-sensor method. Regarding fault detection rate, the proposed method achieves a high detection rate of 99.1%, significantly higher than the 91.1% of the GPS method and the 97.2% of the multi-sensor method. Data fluctuations reflect the impact of sensor drift, battery aging, and environmental interference in the actual engineering environment, verifying the robustness and practicality of the method in complex tunnel environments.

[0091] The analysis of the above simulation results demonstrates the effectiveness of the embodiments of the present invention in solving problems such as untimely fire warnings, low accuracy in fire source location, and the disconnect between emergency response and post-event analysis.

[0092] Specifically, the digital twin-based positioning method significantly improves the accuracy of fire source location, solving the problem of low positioning accuracy; the integrated UAV inspection-intervention decision-making method effectively addresses the issue of untimely fire warnings by demonstrating excellent performance in detection accuracy and response time; and the high-precision environmental reconstruction capability of the data fusion model provides a unified data foundation for emergency response and post-event analysis, bridging the gap between the two. These innovative solutions collectively constitute a comprehensive and efficient fire prevention and control system for cable tunnels, significantly improving the safety operation and maintenance level of cable tunnels.

[0093] Embodiment 2 of the present invention provides a cable tunnel unmanned aerial vehicle (UAV) firefighting system that combines emergency intervention and inspection early warning, comprising: The data acquisition module is used for the UAV to enter the cable tunnel and collect environmental point cloud data, UAV motion status data and two-dimensional thermal imaging frame data in real time through airborne sensors; The digital twin module is used to build and update a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel based on real-time collected data. The anomaly detection module is used to locate abnormal heat sources based on a thermal anomaly function, building upon a digital twin model. After locating the abnormal heat source, it calculates a comprehensive early warning index. ,like If the preset threshold is exceeded, an emergency fire is triggered, and the emergency intervention module is activated; otherwise, the inspection and early warning module is activated. The emergency intervention module is used to analyze abnormal heat sources and correct trajectories based on a digital twin model, guiding fire extinguishing devices to carry out targeted elimination operations. The inspection and early warning module is used to analyze the temperature rise trend and identify potential hidden dangers by comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, so as to issue early warnings. The review and optimization module is used to archive and store key data, model snapshots, and result reports generated throughout the process, for post-event review analysis and model optimization.

[0094] Embodiment 3 of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to perform operations according to the instructions to execute the steps of the method.

[0095] Embodiment 4 of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0096] Compared with the prior art, the beneficial effects of the present invention include at least the following: This invention constructs and updates a digital twin model containing the three-dimensional geometric structure and thermal field distribution of cable tunnels based on real-time collected data. This spatiotemporal digital twin model breaks through the limitations of traditional fixed sensor networks, realizing comprehensive perception and dynamic monitoring of the cable tunnel environment. By fusing multi-source sensor data such as lidar, inertial measurement units, and infrared thermal imagers, a high-precision three-dimensional geometric environment model and spatiotemporal dynamic projection of the temperature field are constructed, providing a solid data foundation for accurate fire source location and fire trend analysis. This not only significantly improves the timeliness of fire early warning but also greatly enhances the accuracy of fire source location, effectively solving the problems of untimely early warning and inaccurate location in existing technologies.

[0097] This invention proposes a fire source location algorithm based on thermal anomalies. It integrates local anomalies, global temperature rise rate, and fire probability models into a unified comprehensive early warning index. Fire situation analysis is performed based on this comprehensive early warning index, which solves the problem of unreliability of single threshold judgment, realizes multi-dimensional comprehensive decision-making, and provides accurate target information for subsequent fire fighting operations.

[0098] This invention achieves closed-loop management across the entire chain, from hazard warning and precise intervention to post-event analysis, through a digital twin engine, autonomous navigation, and fire suppression execution. This not only improves the fire prevention and control capabilities of cable tunnels but also significantly enhances the efficiency and reliability of safe operation and maintenance, realizing a comprehensive improvement from hazard warning and precise intervention to post-event analysis.

[0099] The dual-modal workflow and optimized control strategy proposed in this invention effectively bridge the technical gap between emergency response and post-event analysis, significantly improving the overall response efficiency and intelligence level of the system. This invention integrates fire combustion dynamics, heat conduction equations, extinguishing agent effects, and UAV control inputs into a unified dual-modal coupled dynamics framework, establishing a joint dynamic model of UAV status and cable tunnel thermodynamics. This lays the foundation for achieving coordinated optimal control of UAV trajectory and firefighting mission, significantly improving the accuracy of emergency intervention and system efficiency.

[0100] This invention achieves integrated decision-making for UAV inspection and intervention based on spatiotemporal digital twins. The flexibility and maneuverability of the UAV platform and the real-time perception capabilities of digital twin technology offer significant advantages in coverage and dynamic response compared to solutions using fixed sensor networks. Compared to solutions primarily based on track-based mobile robots, this invention, by introducing autonomous navigation in three-dimensional space and precise fire suppression control, demonstrates superior performance in system response speed and fire source handling efficiency. Particularly in tunnel environments where GPS signals are obstructed, high-precision three-dimensional positioning and map construction are achieved by fusing laser point cloud and inertial navigation data, overcoming the challenge of autonomous navigation for UAVs in complex cable tunnel environments and solving the problem of insufficient navigation and positioning accuracy in complex environments with existing technologies.

[0101] This invention can be widely applied to fire prevention and control of critical infrastructure such as cable tunnels and underground utility tunnels. It can significantly improve the timeliness of fire early warning, increase the accuracy of fire source location, and optimize the efficiency of emergency response. Given its excellent performance in complex environments and comprehensive closed-loop management capabilities, this invention has good application prospects in the safety management of smart city infrastructure.

[0102] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0103] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0104] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0105] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and inspection / early warning, characterized in that: include: S1: The drone enters the cable tunnel and collects environmental point cloud data, drone motion status data and two-dimensional thermal imaging frame data in real time through airborne sensors; S2: Construct and update a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel based on real-time collected data; S3: Based on the digital twin model, locate the abnormal heat source based on the thermal anomaly function. After locating the abnormal heat source, calculate the comprehensive early warning index. If the comprehensive early warning index exceeds the preset threshold, trigger an emergency fire and proceed to S4; otherwise, proceed to S5. S4: Based on the digital twin model, perform abnormal heat source analysis and ballistic correction to guide the fire extinguishing device to carry out targeted elimination operations, then proceed to S6; S5: By comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, analyze the temperature rise trend and identify potential hidden dangers for early warning; S6: Archives key data, model snapshots, and results reports generated throughout the process for post-event review and model optimization.

2. The method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and inspection / early warning, as described in claim 1, is characterized in that: In S2, a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel is constructed and updated based on real-time collected data, specifically including: S201: Perform geometric and pose calculations on the real-time collected environmental point cloud data and the motion state data of the UAV to generate a three-dimensional geometric point cloud map of the cable tunnel and the real-time pose of the UAV in the map. S202: Process the real-time acquired two-dimensional thermal imaging frame data, assign a corresponding temperature value to each pixel, and form a two-dimensional temperature matrix. S203: Based on the extrinsic calibration matrix between sensors and the real-time pose of the UAV in the map, calculate the projection relationship of each frame of thermal imaging in three-dimensional space; S204: Project the two-dimensional temperature matrix onto the three-dimensional geometric point cloud map according to the projection relationship, and assign a corresponding temperature attribute value to each three-dimensional point in the three-dimensional geometric point cloud map to form a digital twin model of the current moment containing the three-dimensional geometric structure and thermal field distribution of the cable tunnel.

3. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and early warning systems, as described in claim 1, is characterized in that: In S3, the location of abnormal heat sources is determined based on the thermal anomaly function, as detailed below: in, Location of abnormal heat source; This is a function for thermal anomaly. The set of all points in the current digital twin model; For the temperature field in the digital twin model, and These represent the mean and standard deviation of historical temperature data, respectively.

4. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and early warning systems, as described in claim 1, is characterized in that: In S3, the comprehensive early warning index is: in, , , These are the weighting coefficients; For temperature field and insulation degradation status The fire probability model; F is a logical variable representing a fire event, with a value of 1 indicating a fire has occurred and a value of 0 indicating no fire has occurred; This represents the thermal anomaly function value corresponding to the location of the abnormal heat source. This is a function for thermal anomaly. This represents a volume element.

5. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and early warning systems, as described in claim 1, is characterized in that: In S4, abnormal heat sources are analyzed and ballistics are corrected based on a digital twin model to guide the fire extinguishing device to carry out targeted elimination operations, specifically including: S401: Upon receiving an emergency fire trigger signal, retrieve the latest digital twin model of the moment the emergency fire was triggered, and based on the comprehensive early warning index... Filter the point cloud in the model to select all comprehensive early warning indices. Points exceeding a preset threshold form an abnormal heat source cloud cluster; S402: Calculate the geometric center coordinates of the abnormal heat source point cloud cluster in three-dimensional space as the target coordinates, and obtain the real-time pose information of the UAV and the relative attitude of the fire extinguishing device. S403: The ballistic correction model calculates the projection direction compensation required to hit the target based on the target coordinates, the real-time pose of the UAV, the relative attitude of the fire extinguishing device, and the environmental parameters obtained by the airborne sensors. This includes the pitch angle correction α and the azimuth angle correction β. S404: Correct the target direction according to the projection direction compensation amount, and adjust the projection axis of the drone and the fire extinguishing device to the corrected target direction; S405: The drone performs the deployment of the airborne fire extinguishing device. After deployment, it continuously monitors the temperature changes in the target area to determine whether the fire has been effectively suppressed.

6. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and early warning systems, as described in claim 1, is characterized in that: In S5, by comparing the current time digital twin model with the historical inspection cycle digital twin model, the temperature rise trend is analyzed and potential hidden dangers are identified for early warning, specifically including: S501: The UAV maintains a stable flight state according to the preset trajectory, performs daily inspection tasks, and obtains the digital twin model Model-t0 of the current time and the digital twin model Model-t1 of the previous inspection cycle; S502: Employs a point cloud registration algorithm to spatially align Model-t0 and Model-t1; S503: Perform a difference operation on the aligned Model-t0 and Model-t1 to obtain the temperature change at each location point in the tunnel; S504: Determine whether the temperature change exceeds the sensitivity threshold of the temperature rise trend. If so, the corresponding location point is taken as a potential hidden danger point, and a digital twin model of an earlier inspection cycle is obtained to construct a temperature change time series of the potential hidden danger point. S505: Perform fitting analysis on the temperature change time series, calculate the future temperature rise rate, and predict the time required to reach the alarm temperature threshold based on the rate. S506: Graded early warning based on prediction results: If the predicted time is lower than the set value, a high-priority early warning is generated; otherwise, a general attention notice is generated, and the monitoring frequency of the area corresponding to the potential hazard point is automatically increased in future inspections.

7. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and inspection / early warning, as described in claim 1, is characterized in that: The UAV adopts the following optimized control strategy: A joint dynamic model based on UAV status and cable tunnel thermodynamics And the definition of Hamiltonian function for the optimal control problem of unmanned aerial vehicles. According to the Hamiltonian function To obtain the optimal control input for achieving coordinated optimization of state trajectory and control input. The details are as follows: in, Let be the system state vector. To control the input vector, Let t be the accompanying variable and t be the time. This represents the immediate cost obtained based on the optimal control problem of unmanned aerial vehicles (UAVs). The weighted quadratic norm representing the state deviation. This represents the energy consumption or risk weight of the control input.

8. A method for extinguishing fires in cable tunnels using unmanned aerial vehicles (UAVs) that combines emergency intervention and early warning systems, as described in claim 7, is characterized in that: A joint kinetic model of UAV status and cable tunnel thermodynamics Specifically as follows: in, For fire combustion dynamics response; The thermodynamic enthalpy per unit volume or per unit mass of the composite medium in the cable tunnel; This represents the current temperature field inside the tunnel. For system enthalpy; These are insulation degradation parameters; The ambient reference temperature This is the heat conduction term for the temperature field; It is a function relating resistance to temperature; Let be the mass flow rate of the extinguishing agent injected at time t; Δ is the enthalpy change of the reaction between the extinguishing agent and combustion; Q is the heat release rate of the fire source. Let be the weight coefficient of the i-th sub-control module; Let i be the state variable of the i-th subsystem in the joint dynamics model. Let i be the dynamic response function of the i-th subsystem; These are physical and control parameters.

9. A cable tunnel unmanned aerial vehicle (UAV) firefighting system that combines emergency intervention and inspection early warning, operating the method described in any one of claims 1-8, characterized in that, The system includes: The data acquisition module is used for the UAV to enter the cable tunnel and collect environmental point cloud data, UAV motion status data and two-dimensional thermal imaging frame data in real time through airborne sensors; The digital twin module is used to build and update a digital twin model containing the three-dimensional geometry and thermal field distribution of the cable tunnel based on real-time collected data. The anomaly detection module is used to locate the location of abnormal heat sources based on the thermal anomaly function on the basis of the digital twin model. After locating the abnormal heat source, it calculates the comprehensive early warning index. If the comprehensive early warning index exceeds the preset threshold, it triggers an emergency fire and enters the emergency intervention module; otherwise, it inspects the early warning module. The emergency intervention module is used to analyze abnormal heat sources and correct trajectories based on a digital twin model, guiding fire extinguishing devices to carry out targeted elimination operations. The inspection and early warning module is used to analyze the temperature rise trend and identify potential hidden dangers by comparing the digital twin model of the current time with the digital twin model of the historical inspection cycle, so as to issue early warnings. The review and optimization module is used to archive and store key data, model snapshots, and result reports generated throughout the process, for post-event review analysis and model optimization.

10. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-8.

Images