Fire sensing method of double-layer cable tunnel structure and related device

Abstract

The invention belongs to the technical field of fire sensing and detection, and provides a fire sensing method and related device for a double-layer cable tunnel structure, and the method comprises the steps: respectively collecting the temperature, smoke and image data of a double-layer cable tunnel, and carrying out the double-layer data linkage; respectively constructing a temperature distribution model, a smoke diffusion model and an image target detection model according to the collected temperature, smoke and image data; and performing temperature, smoke and image multi-modal data fusion based on the temperature distribution model, the smoke diffusion model and the image target detection model, and when the fused data meets a set fire threshold, determining that a fire occurs. Through the fire sensing and verification method aiming at the characteristics of the double-layer cable tunnel, scientific and reliable decision support is provided for improving tunnel fire safety management, the safety of personnel and equipment in the tunnel is ensured, and meanwhile, the standardization process of the fire sensing technology is promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of fire perception and detection, and in particular relates to a fire perception method and related devices for a double-layer cable tunnel structure. Background Art

[0002] Urban underground utility corridors and cable tunnels have gradually become an important part of the power system. Among them, double-layer cable tunnels are increasingly used in cable laying in large cities and industrial parks due to their high space utilization efficiency and good safety performance. However, the fire safety issues brought about by this double-layer structure are particularly prominent, and fire perception and early warning have become one of the difficulties in cable tunnel safety management.

[0003] At present, the technologies for double-layer cable tunnel fire perception mainly focus on the following aspects: (1) Single parameter detection technology: Traditional fire detection systems mostly use a single parameter, such as temperature or smoke sensors. These systems have a certain detection effect in single-layer cable tunnels, but in double-layer structures, due to the complex spread of fire and the diversity of spatial environment, there are often problems of low detection accuracy and high false alarm rate. In particular, single parameter detection has poor sensitivity to small-scale combustion in the early stages of a fire, making it difficult to provide timely warnings.

[0004] (2) Multi-parameter fire detection technology: In recent years, with the development of sensor technology, multi-parameter fire detection has gradually become a research hotspot. Multi-parameter fire detection can more effectively improve the accuracy of fire detection through the fusion analysis of temperature, smoke, optical signals and other data. However, the current multi-parameter technology is mainly used in open spaces or single-layer structure environments, and there is a lack of special research on the complex environment of double-layer cable tunnels.

[0005] (3) Fire image recognition technology: Fire recognition technology based on image processing and artificial intelligence is also developing rapidly. Using thermal imaging and visible light cameras to identify and locate fire sources in real time can intuitively perceive fire conditions. However, this type of technology is limited by smoke obstruction and image noise interference in complex environments, and its application in cable tunnel environments still faces many challenges.

[0006] Existing problems and challenges: (1) Complex fire spread characteristics: The air flow and heat conduction in the double-layer cable tunnel are significantly affected by the tunnel structure. The fire may spread rapidly to the upper and lower layers along the cables or building structures. The smoke flow path is difficult to predict, making it difficult for traditional detection methods to capture the early characteristics of the fire.

[0007] (2) Serious false alarms and missed alarms: Due to the variability of the tunnel environment, such as humidity, wind speed, air flow and other factors, the fire detection system is prone to false alarms or missed alarms, which affects the actual application effect of the system. In addition, the fire detection needs at different levels are different, and it is difficult to cover all fire scenarios with a single detection method. Summary of the invention

[0008] The purpose of the present invention is to provide a fire sensing method and related devices for a double-layer cable tunnel structure, so as to solve the problem that the traditional solution has inaccurate fire feature capture and many false alarms or missed alarms.

[0009] To achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a fire sensing method for a double-layer cable tunnel structure, comprising: Collect temperature, smoke and image data of double-layer cable tunnels respectively, and link the double-layer data; According to the collected temperature, smoke and image data, a temperature distribution model, a smoke diffusion model and an image target detection model are constructed respectively; Multimodal data fusion of temperature, smoke and image is performed based on the temperature distribution model, smoke diffusion model and image target detection model. When the fused data meets the set fire threshold, the fire is determined to have occurred.

[0010] Furthermore, the collecting of temperature, smoke and image data of the double-layer cable tunnel includes: Temperature detectors are arranged at the top ceiling of each upper tunnel section, at the intersection nodes, corners and branches of the upper tunnel, and in the area where cables are densely distributed in the upper tunnel. The distributed temperature detectors can monitor the accumulation and conduction of heat in real time. Laser smoke detectors are arranged at each smoke exhaust port, as well as at the entrance and exit of the upper tunnel to monitor the spread of fire smoke. Thermal imaging and visible light camera combination devices are arranged at intervals in each tunnel section. Thermal imaging is used to monitor temperature anomalies, and visible light cameras are used to identify smoke and flames. Laser smoke detectors are arranged in the lower tunnels near the tunnel side walls or at low intervals along both sides of the tunnel; Temperature detectors are arranged at intervals on each cable tray in the lower tunnel and on the equipment boxes to monitor heat accumulation and conduction in real time; visible light cameras are arranged at each tunnel exit and emergency evacuation channel to monitor the evacuation path of personnel; thermal imaging cameras are arranged in the cable tray area to monitor thermal anomalies of the cables in real time.

[0011] Furthermore, the two-layer data linkage includes: All detector signals are transmitted to the monitoring center via wired or wireless networks. The monitoring center adjusts the sensitivity of the detectors in real time based on the detector data. Establish a data linkage mechanism between the two layers: according to any abnormal data in the temperature, smoke and image data of any layer, adjust the sensitivity of the detector on the other layer and enter the high-sensitivity mode.

[0012] Furthermore, the building of the temperature distribution model includes: Deploy a fiber optic temperature sensor network in the double-layer tunnel, using time division multiplexing (TDM) for data transmission; collect temperature distribution data along the length of the tunnel in real time, sense temperature changes through optical fiber, and use the Raman scattering principle to accurately locate the temperature change point; set multiple temperature alarm thresholds, and when the temperature reaches the threshold, the system triggers different levels of warnings based on the temperature change rate; The long short-term memory network (LSTM) is used to process historical temperature data and analyze the temperature change trend. According to the prediction results output by the LSTM model, the temperature alarm threshold is dynamically adjusted. The support vector machine (SVM) is used to classify the temperature data and distinguish between normal fluctuations and signs of fire.

[0013] Furthermore, the smoke diffusion model is constructed, including: The detector uses the principle of laser scattering to detect fine smoke particles, and uses an optical sensor array to improve detection sensitivity; an adaptive filtering algorithm is used to eliminate dust and water vapor interference from non-fire smoke; and by analyzing the rate of change of smoke concentration, it is determined whether the fire is fast-spreading or slow-developing. Combined with the Navier-Stokes equations of fluid mechanics, a smoke diffusion model is established. The installation position of the smoke detector is optimized according to the model results. The smoke concentration data collected by the sensor is compared with the model prediction value in real time, and the least squares method is used to correct the abnormal smoke concentration detection results.

[0014] Furthermore, the building of the image target detection model includes: Data is transmitted via the serial port or Ethernet, and convolutional neural network (CNN) is used for image processing. The model is trained to identify flame morphology and smoke characteristics. The model is trained using an image dataset containing various fire scenes. The YOLO real-time target detection algorithm is combined to identify smoke and open flames.

[0015] Furthermore, the multimodal data fusion of temperature, smoke and image is performed based on the temperature distribution model, the smoke diffusion model and the image target detection model, and when the fused data meets the set fire threshold, the fire is determined to have occurred, including: The Bayesian network algorithm is used to fuse image data with temperature and smoke signals. The confidence level of each parameter is calculated and the threshold for fire determination is set. When all signals meet the fire characteristics at the same time, a fire is determined to have occurred. The random forest model is used to train historical fire data, and the sensitivity parameters of the detector are automatically corrected according to the real-time monitoring data. When a sensor sends a fire signal, multiple parameters are called for verification, and the temperature, smoke and image features are integrated. The decision tree algorithm is used to check the data consistency. If all signals are consistent, a fire alarm is triggered. If the signals are inconsistent, a deep analysis process is entered, and anomaly detection is performed through a clustering algorithm.

[0016] In a second aspect, the present invention provides a fire sensing system for a double-layer cable tunnel structure, comprising: The data acquisition module is used to collect temperature, smoke and image data of the double-layer cable tunnel respectively and to link the double-layer data; A model building module is used to build a temperature distribution model, a smoke diffusion model and an image target detection model according to the collected temperature, smoke and image data; The fusion discrimination module is used to fuse multimodal data of temperature, smoke and image based on the temperature distribution model, smoke diffusion model and image target detection model. When the fused data meets the set fire threshold, it is determined that a fire has occurred.

[0017] In a third aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the fire perception method for the double-layer cable tunnel structure when executing the computer program.

[0018] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the fire sensing method for the double-layer cable tunnel structure.

[0019] Compared with the prior art, the present invention has the following technical effects: The present invention aims at the structural characteristics of the double-layer cable tunnel and takes into account complex factors such as air flow and heat conduction, so that the fire perception method is more suitable for actual application scenarios and improves the sensitivity and accuracy of the early warning.

[0020] By placing temperature detectors, laser smoke detectors and cameras at key locations in the upper and lower tunnels, comprehensive monitoring of the tunnel environment is achieved to ensure that fires can be discovered in a timely manner.

[0021] The system integrates multiple monitoring parameters such as temperature, smoke, and images to form a diversified means of fire perception. This comprehensive monitoring method can significantly improve the accuracy of fire identification and reduce the false alarm rate caused by false alarms of a single parameter. Multimodal data fusion technology is used to conduct a comprehensive analysis of temperature, smoke, and image data, further improving the reliability and accuracy of fire identification.

[0022] The present invention provides real-time safety monitoring and early warning for operation and maintenance personnel, so that fires can be discovered and handled in the first place, thereby greatly reducing the risks and losses caused by fires. The efficient fire perception method provides a scientific and reliable basis for decision-making, helping operation and maintenance personnel to make more accurate and timely decisions.

[0023] This invention promotes the standardization of fire sensing methods through technological innovation and provides the industry with unified detection standards and specifications. This helps to improve the overall safety level of the industry and promote the widespread application and promotion of fire sensing technology. The standardized fire sensing method can provide reference and reference for other companies and institutions in the industry, and promote the entire industry to develop in a safer and more reliable direction.

[0024] Through multi-parameter monitoring and data fusion technology, the accuracy and reliability of fire identification have been significantly improved. Real-time safety monitoring and early warning enable operation and maintenance personnel to respond to fire incidents quickly, reducing the risks and losses caused by fire.

[0025] In summary, this technical solution provides scientific and reliable decision-making support for improving tunnel fire safety management through fire perception and verification methods tailored to the characteristics of double-layer cable tunnels, ensures the safety of personnel and equipment in the tunnel, and promotes the standardization process of fire perception technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flow chart of the present invention.

[0027] Figure 2 This is a framework diagram of the fire sensing technology based on the structural features of the double-layer cable tunnel of the present invention. DETAILED DESCRIPTION

[0028] The present invention is further described below in conjunction with the accompanying drawings: Example 1, please refer to Figure 1 The present invention provides a fire sensing method for a double-layer cable tunnel structure, comprising: Collect temperature, smoke and image data of double-layer cable tunnels respectively, and link the double-layer data; According to the collected temperature, smoke and image data, a temperature distribution model, a smoke diffusion model and an image target detection model are constructed respectively; Multimodal data fusion of temperature, smoke and image is performed based on the temperature distribution model, smoke diffusion model and image target detection model. When the fused data meets the set fire threshold, the fire is determined to have occurred.

[0029] The present invention realizes multi-dimensional and comprehensive perception of the tunnel environment by collecting temperature, smoke and image data of the double-layer cable tunnel respectively. This data collection method can capture various subtle changes before the fire occurs, providing rich data support for subsequent fire judgment.

[0030] By establishing a data linkage mechanism between the two layers, the solution realizes real-time sharing and mutual verification of upper and lower tunnel data. When abnormal data appears on any layer, the sensitivity of the detector on the other layer can be quickly adjusted to enter high-sensitivity mode, thereby improving the sensitivity and accuracy of fire perception.

[0031] The present invention constructs a temperature distribution model, a smoke diffusion model and an image target detection model based on the collected data. These models can accurately reflect the changing patterns of temperature, smoke and images in the tunnel, providing a scientific basis for fire judgment. Through the in-depth analysis of the data by the model, the scheme can dig out potential features before the fire occurs, such as abnormal temperature rise, smoke concentration changes, etc. This deep data analysis helps to improve the accuracy and timeliness of fire identification.

[0032] The present invention uses multimodal data fusion technology to comprehensively analyze temperature, smoke and image data. This fusion method can fully utilize the advantages of various data and improve the reliability and accuracy of fire identification. By setting the fire threshold, the solution can intelligently determine whether a fire has occurred. When the fused data meets the set threshold, the fire alarm is triggered, thereby achieving timely response and processing of the fire.

[0033] Through efficient data collection, accurate model building and intelligent fire judgment, the solution significantly improves the fire perception capability of double-layer cable tunnels. This helps operation and maintenance personnel to detect and handle fire incidents in a timely manner, ensuring the safety of personnel and equipment in the tunnel. The application of multimodal data fusion and double-layer data linkage mechanism enhances the reliability and stability of the system. Even if some sensors fail or misreport, the system can still accurately judge the fire situation and avoid false alarms and missed alarms.

[0034] In summary, this technical solution significantly improves the fire perception capability and safety management level of double-layer cable tunnels through efficient data collection, accurate model construction, intelligent fire judgment, multi-modal data fusion and double-layer data linkage mechanism. This provides a strong guarantee for the safety of personnel and equipment in the tunnel, and also provides a useful reference for the further development and application of fire perception technology.

[0035] Embodiment 2, the present invention provides a fire sensing method for a double-layer cable tunnel structure, specifically comprising: 1. Fire perception method based on the structural characteristics of double-layer tunnels The fire spread law in a double-layer cable tunnel is limited by the influence of the upper and lower layers. The heat transfer and smoke diffusion characteristics of the fire between the two layers are completely different. Therefore, the fire detection system must be able to respond to the fire characteristics of the upper and lower layers respectively.

[0036] 1. Layered detector layout and optimization In view of the structural characteristics and fire spreading rules of double-layer cable tunnels, the present invention adopts a layered and focused coverage strategy in the arrangement of detectors. The specific details are as follows: (1) Arrangement of detection points in the upper tunnel The upper tunnel is usually the path for heat and flames to spread upward, so the temperature change is sensitive and spreads quickly. The detector layout is as follows: Temperature detector layout: A high-sensitivity temperature detector is arranged every 15-20 meters at the top ceiling of each tunnel. The accumulation and conduction of heat are monitored in real time through densely distributed temperature detection points. Additional temperature detectors are arranged at the intersection nodes, corners and branches of the upper tunnel to monitor the changes in the heat flow field. These locations are key areas where fire heat is easy to accumulate and spread. Temperature detectors are installed in areas with dense cable distribution (such as main cable racks) in the upper tunnel to ensure that fire signals in high fire risk areas can be captured as soon as possible.

[0037] Smoke detector arrangement: A laser smoke detector is arranged at each smoke exhaust port to monitor the spread of fire smoke. The smoke detector can timely sense whether the smoke has reached the vicinity of the smoke exhaust system, thereby commanding the smoke exhaust system to start. Smoke detectors are arranged at the entrance and exit of the upper tunnel, so that changes in the external air and internal environment can be sensed immediately, as well as whether the fire smoke has spread to the entrance and exit of the tunnel.

[0038] Image monitoring point arrangement: A combination of thermal imaging and visible light cameras is arranged every 30 meters in each tunnel section to provide panoramic monitoring of the overall condition of the tunnel. Thermal imaging is used to monitor temperature anomalies, and visible light cameras are used to identify smoke and flames. Additional image monitoring devices are arranged in locations such as cable equipment boxes and substation equipment. Once a fire occurs in these areas, it may lead to more serious consequences, so special monitoring is required.

[0039] (2) Arrangement of detection points in the lower tunnel The fire characteristics of the lower tunnel are more significant smoke accumulation, and it may be difficult to quickly rise to high temperature in the early stage of the fire. Therefore, the layout of the detection points in the lower tunnel focuses on smoke monitoring and accurate temperature detection: Laser smoke detector placement: Laser smoke detectors are placed every 10-15 meters, close to the tunnel side walls or installed at low positions along both sides of the tunnel. This arrangement ensures that smoke is detected quickly when it is on the ground or near the cable rack, especially capturing smoke signals in the early stages of a fire. Additional smoke detectors are placed in low-lying areas in the tunnel, such as near low-lying drainage ditches, as smoke tends to accumulate in these areas.

[0040] Temperature detector arrangement: Temperature detectors are arranged on each cable tray, with an interval of no more than 10 meters, especially along the densely laid cable areas. The arrangement focuses on monitoring possible short circuits or burning points of the cables. Since the fire characteristics of the lower tunnel are usually characterized by slow heat accumulation and local temperature rise, high-precision temperature detectors are added on the top of the tunnel and near important equipment boxes.

[0041] Image monitoring point layout: Visible light cameras are placed at each tunnel exit and emergency evacuation channel to ensure that the fire can be monitored to see if it threatens the evacuation route. Thermal imaging cameras are placed to cover the entire cable tray area to monitor the thermal anomalies of the cables in real time to ensure that the fire point and the direction of fire spread can be accurately located.

[0042] (3) Overall linkage of detector layout Integrated monitoring center: All detector signals are transmitted to the integrated monitoring center via wired or wireless networks. The monitoring center is equipped with a data processing server to analyze the data of each detector in real time and use multi-parameter fusion algorithms for fire identification and alarm.

[0043] Adaptive sensitivity adjustment: The sensitivity of the detector will be adaptively adjusted according to different time periods and environmental conditions. For example, at night when the cable load is low, the threshold of the temperature detector can be automatically lowered to improve the early detection capability of fire.

[0044] 2. Two-way linkage response mechanism Data linkage between upper and lower layers: Through the data linkage mechanism between the two layers, when one layer detects fire characteristics, the detectors on the other layer will automatically enter high-sensitivity mode to respond to fire changes more quickly and prevent the fire from spreading to the other layer without being detected. Data is transmitted to the control center via wireless or wired networks for real-time analysis.

[0045] Intelligent alarm and zoning management: After detecting a fire, the system will automatically activate the corresponding emergency plan based on the fire location, including starting the smoke exhaust system, closing the vents, and sending real-time reports on the fire location and spread to emergency personnel. The fire response strategy is adjusted according to the tunnel zoning to prevent the fire from spreading.

[0046] 3. Thermal flow field simulation and detector optimization layout CFD (Computational Fluid Dynamics) was used to simulate the thermal flow field of the double-layer tunnel and analyze the spread path of the fire at different locations. Based on the simulation results, the layout and density of the detectors were optimized to ensure no blind spot monitoring, and the response time of the detectors was optimized through heat flow distribution.

[0047] 2. Multi-parameter verification method for fire perception The fire sensing system of the present invention is specifically designed for the characteristics of double-layer tunnel structures and adopts a multi-parameter joint verification method to improve the accuracy and robustness of fire detection. The following is a specific technical solution for each module: (1) Temperature sensing module Distributed fiber temperature sensor (DTS) network: Deploy a fiber temperature sensor network in a double-layer tunnel, using time division multiplexing (TDM) technology for data transmission. Collect temperature distribution data along the length of the tunnel in real time, sense temperature changes through optical fiber, and use Raman scattering to accurately locate temperature change points. Set multiple temperature alarm thresholds (such as 50°C, 70°C, 90°C). When the temperature reaches the threshold, the system triggers different levels of warnings based on the temperature change rate.

[0048] Intelligent temperature analysis algorithm: Use long short-term memory (LSTM) networks to process historical temperature data and analyze temperature change trends. Dynamically adjust the temperature alarm threshold based on the prediction results output by the LSTM model to ensure accuracy in complex environments. Use support vector machines (SVM) to classify temperature data and distinguish between normal fluctuations and signs of fire.

[0049] (2) Smoke sensing module High-sensitivity laser smoke detector: The detector uses the principle of laser scattering to quickly detect fine smoke particles. It uses an optical sensor array to improve detection sensitivity. It uses adaptive filtering algorithms (such as Kalman filtering) to eliminate non-fire smoke interference such as dust and water vapor to improve detection accuracy. By analyzing the rate of change of smoke concentration (such as when the change exceeds a certain set value), it can be determined whether the fire is fast-spreading or slow-developing.

[0050] Smoke diffusion model: Combined with the Navier-Stokes equations of fluid mechanics, a smoke diffusion model is established. The installation location of the smoke detector is optimized based on the model results to ensure that possible smoke sources in the double-layer tunnel are covered. The smoke concentration data collected by the sensor is compared with the model prediction value in real time, and the least squares method is used to correct the abnormal smoke concentration detection results.

[0051] (3) Image perception module Image acquisition and analysis: High-resolution thermal imaging cameras and visible light cameras are configured to transmit data to the central processing unit via serial ports or Ethernet. Convolutional neural networks (CNN) are used for image processing to train models to identify flame morphology and smoke characteristics. The model is trained using image datasets containing various fire scenes. Combined with the YOLO (You Only Look Once) real-time target detection algorithm, the recognition speed and accuracy of fire smoke and open flames are improved.

[0052] Multimodal data fusion: Bayesian network algorithm is used to fuse image data with temperature and smoke signals to achieve comprehensive analysis of multi-parameter data. By calculating the confidence level of each parameter and setting the threshold for fire determination, all signals are required to meet the fire characteristics at the same time, reducing the false alarm rate.

[0053] (4) Intelligent identification and verification mechanism Machine learning aided decision-making: Use the random forest model to train historical fire data and improve the accuracy of the model through feature selection. Automatically calibrate the sensitivity parameters of the detector based on real-time monitoring data to ensure its effectiveness under different environmental conditions.

[0054] Real-time verification and feedback mechanism: When a sensor sends a fire signal, the system automatically calls the multi-parameter verification mechanism to conduct a comprehensive analysis of temperature, smoke and image features. The decision tree algorithm is used to check data consistency. If all signals are consistent, the fire alarm is triggered; if the signals are inconsistent, the in-depth analysis process is entered, and anomaly detection is performed through clustering algorithms (such as K-means) to ensure the accuracy of the alarm.

[0055] Through the specific technical solutions of the above modules, this fire sensing system can effectively meet the fire detection needs of double-layer tunnels and provide a set of efficient, accurate and reliable fire monitoring solutions.

[0056] In order to verify the effectiveness and reliability of the fire sensing system in a double-layer cable tunnel, this paper proposes a verification technology solution based on a field fire experiment. This solution fully considers the particularity of the double-layer tunnel structure and comprehensively evaluates the multi-parameter detection capability of the system by accurately simulating real fire scenes and environmental interference factors. The following is the specific experimental step design: (1) Experimental scene construction Real fire scene simulation: A double-layer cable tunnel matching the actual project is built in the experimental base, with real cables and other combustible materials arranged inside the tunnel, and different fire scenes are simulated. In order to fully test the responsiveness and stability of the fire perception system in the double-layer cable tunnel, the experimental conditions cover a variety of actual situations such as fire types, spread paths, environmental interference, etc., and the following verification scheme is provided: Condition 1: Initial fire source fire simulation Objective: Test the system’s sensitivity and response speed to early fire signals and verify its rapid detection capabilities and low-latency alarm mechanism.

[0057] Fire source arrangement: On the cable tray in the lower level of the tunnel, ignite a small fire source, such as a single insulated cable or a small amount of paper and flammable packaging materials. The flame height is controlled at 0.5-1 meter to simulate the initial fire caused by a common cable short circuit.

[0058] Detector arrangement: Distributed fiber optic temperature sensors are arranged along the bracket with a spacing of 5 meters, and high-sensitivity smoke detectors and high-definition thermal imaging cameras are arranged around the fire source.

[0059] Key points of testing: Observe the temperature rise curve of the temperature sensing module, analyze the response time and false alarm rate of the smoke detector, and use the thermal imaging camera to identify changes in flame and smoke morphology. Synchronously record data to evaluate system latency.

[0060] Environmental parameters: Set a windless or light breeze environment with a humidity of 50%-70% to ensure that there is no external interference with the initial fire signal, and focus on evaluating the detection sensitivity in the early stages of the fire.

[0061] Condition 2: Local high temperature fire spreads Objective: To evaluate the detection capability, positioning accuracy and multi-parameter joint analysis of the system during fire spread.

[0062] Fire source arrangement: Select a cable concentration area to ignite multiple cables, and the flames gradually spread to the adjacent cable brackets to form a local high-temperature fire. Use fire barriers to guide the flames to spread to the upper level of the tunnel to simulate the longitudinal and lateral spread of fire in a double-layer structure.

[0063] Detector layout: Temperature sensors cover both sides of the fire source, and smoke detectors are installed on the upper and lower floors respectively. Thermal imaging cameras monitor the development of the flame.

[0064] Detection focus: Analyze the temperature change rate, smoke diffusion path and flame image characteristics during the fire spread process. Record the response time of each detector to ensure the synchronization of multi-parameter data and the fusion analysis effect. Observe whether the system can accurately locate the fire source and issue an alarm in time when the flame crosses layers.

[0065] Environmental parameters: Set the natural wind speed to 1-3 m / s to simulate the common airflow conditions in the tunnel and verify the accuracy of the smoke detector under airflow disturbance.

[0066] Condition 3: Large-scale fire in the entire tunnel Objective: To verify the system’s responsiveness, data processing capabilities, and overall stability under fire conditions.

[0067] Fire source arrangement: Multiple fire source points are evenly distributed along the entire length of the tunnel, including the cable accumulation area on the upper and lower brackets, forming a large-scale tunnel-wide fire. The fire source fuel is selected from high-temperature released industrial oils and plastic cable materials to ensure the generation of thick smoke and high temperature environment.

[0068] Detector layout: Dense temperature and smoke detectors are set up throughout the tunnel, and multiple high-definition image monitoring points are arranged along the fire source path to comprehensively monitor the evolution of the fire.

[0069] Detection focus: The system needs to quickly identify the phenomenon of multiple fires merging and analyze the behavior of flames and smoke under high temperature conditions. Focus on the false alarm and missed alarm rates of the system under extreme conditions, and the stability of the detection data fusion algorithm in large-scale fires. Evaluate the matching degree between the speed of fire spread and the system response time.

[0070] Environmental parameters: Turn off some ventilation equipment and maintain extreme conditions of high humidity and high smoke concentration to simulate a fire scene in a closed tunnel.

[0071] Case 4: Wind speed disturbance simulation Objective: To test the detection accuracy of the system in strong wind conditions, especially the smoke diffusion speed and flame image recognition.

[0072] Fire source arrangement: Ignite the fire source in the middle of the lower level of the tunnel near the air duct to burn highly volatile oils and plastic materials, producing a large amount of easily diffused thick smoke to simulate the impact of airflow in the tunnel during a fire.

[0073] Detector placement: Smoke detectors and image monitoring equipment are placed near and away from the air duct. Temperature sensors extend along the path of the fire source to capture temperature fluctuations under the influence of high wind speed.

[0074] Key points of the test: Observe the response speed and concentration change curve of the smoke sensing module to verify the accuracy of the smoke model under wind speed interference. The image sensing module needs to identify the flame stability under strong wind conditions to ensure that the fire characteristics will not be misidentified due to airflow interference.

[0075] Environmental parameters: The wind speed was gradually adjusted from 2 m / s to 15 m / s to simulate airflow conditions of different intensities and to examine the stability and adaptability of the system.

[0076] Case 5: Non-fire event simulation Objective: To verify the system's ability to resist false alarms and test its performance under non-fire interference conditions.

[0077] Interference source arrangement: Set up non-fire interference sources in the tunnel, such as dust generated by mechanical operations, water mist sprayed by sprinkler equipment, and high-concentration moisture emitted by vehicle exhaust. Detector arrangement: Install smoke detectors and image monitoring equipment near and far from the interference source. The temperature sensing module needs to monitor the temperature changes in the tunnel in real time to ensure that interference does not trigger an alarm.

[0078] Key points of the test: Observe whether the system can eliminate interference from non-fire events through a multi-parameter verification mechanism, and analyze the sensor's response to high humidity or dusty environments. Verify whether the smoke detector will give a false alarm, and evaluate the image analysis module's ability to identify flames under the interference of water mist.

[0079] Environmental parameters: Increase the humidity to above 90% to simulate watering and humid environment, and examine the system's anti-interference performance to extreme humidity.

[0080] Condition 6: Fire on the upper floor triggers flame backflow on the lower floor Objective: To analyze the nonlinear propagation path of fire in a double-layer tunnel structure and investigate the system's ability to detect flame backflow phenomena.

[0081] Fire source arrangement: multiple fire sources are arranged in the upper tunnel. After ignition, the hot air flow and smoke are guided to spread along the tunnel structure to the end. Then the air flow is designed to flow back to the lower tunnel to simulate the complex flow of flames and hot air flow.

[0082] Detector layout: Temperature sensors and smoke detectors are densely arranged in the reflow area to observe their detection effect on the reflow of hot air. Image perception modules are used to capture the details of the reversal of flame direction to ensure that the system can accurately identify changes in flame characteristics.

[0083] Detection focus: Record the temperature change curve of the flame spreading from the upper layer to the lower layer, analyze the fluctuation of smoke concentration, and verify the system's alarm capability and data fusion accuracy in complex fire paths.

[0084] Environmental parameters: Adjust the wind speed and air pressure in the tunnel to simulate the effect of heat flow reversal in the real environment and examine the adaptability of the system under non-linear flame propagation.

[0085] Condition 7: Multiple fires and combined effects Objective: To test whether the system can accurately detect and promptly alarm when multiple fires occur simultaneously or when fire sources merge.

[0086] Fire source arrangement: Select different locations in the tunnel and ignite multiple small-scale fire sources, such as cable accumulation areas and scattered oil fuels, to observe whether the fire sources will quickly merge into a large fire, causing a sudden increase in temperature and smoke concentration.

[0087] Detector layout: Multi-layer distributed smoke detectors and temperature sensing equipment to monitor the temperature rise rate and smoke diffusion characteristics of each point during the fire source merging process.

[0088] Detection focus: Analyze the system's alarm effect on multiple fires, verify whether the image analysis module can accurately identify the fire development trend when flames merge, optimize the detector layout and multi-parameter fusion algorithm, and ensure that the system can still accurately alarm when fires merge.

[0089] Environmental parameters: set intermittent airflow disturbances to simulate the effects of natural ventilation in the tunnel and external interference on fire fusion, and further test the stability of the system.

[0090] Simulation of environmental interference factors: Airflow and Ventilation: Set different wind speeds and ventilation conditions to reflect common air flow characteristics in tunnels and evaluate the accuracy and response speed of the smoke sensing module.

[0091] Humidity and dust interference: simulate high humidity and dust environment to test the system's false alarm handling ability and ensure that the sensor can still operate stably in complex environments.

[0092] Condition 8: Long-term concealed combustion of small fire sources Objective: To evaluate the detection capability of the system in a long-term concealed combustion environment, especially the sensitivity and stability for the concealed state in the early stage of fire.

[0093] Fire source arrangement: Place a single-strand cable with an insulating sheath in the lower cable tray of the tunnel. Slowly burn the cable material to simulate a hidden fire source. The flame produces weak heat and a small amount of smoke, and the duration is set to more than 1 hour.

[0094] Detector layout: Install high-precision distributed fiber optic temperature sensors and high-sensitivity smoke detectors around the hidden fire source area. The image monitoring system focuses on capturing weak smoke and temperature anomalies.

[0095] Key points of testing: Record the temperature curve as it gradually increases, verify whether the system can identify signs of fire when the temperature fluctuates slightly, and analyze the performance of the detector when the smoke concentration is extremely low. Check whether the system will produce false alarms or missed alarms due to background noise during long-term monitoring, and ensure that the detector can maintain high sensitivity.

[0096] Environmental parameters: Set a constant temperature and humidity environment, maintain low wind speed conditions, simulate the stability of the hidden combustion environment of the fire source, and focus on examining the detection accuracy of the system when there is no obvious environmental disturbance.

[0097] Condition 9: Smoke barrier under local high temperature Objective: To verify the system's detection capabilities in special situations of localized high temperatures and smoke obstruction, especially in double-layer structures with uneven smoke distribution.

[0098] Fire source arrangement: Ignite a high-temperature fire source in the upper tunnel, use fireproof partitions to block the diffusion path of smoke, and only allow part of the smoke to flow into the lower tunnel. Simulate the uneven distribution of smoke in a double-layer tunnel fire.

[0099] Detector arrangement: Temperature sensors are set up around the fire source area to monitor the high temperature area, and smoke detectors are distributed on the upper and lower floors. The image monitoring system focuses on identifying signs of fire when the smoke concentration is blocked.

[0100] Testing focus: Analyze whether the temperature sensing module can independently determine the severity of the fire and the response of the smoke detector when the concentration decreases. Verify whether the system can maintain accuracy when smoke detection is obstructed by combining temperature and image data.

[0101] Environmental parameters: simulate different smoke blockage scenarios, adjust the wind speed to examine the adaptability of the system after smoke blockage, and ensure that the fire sensing system has the ability to detect non-uniform smoke distribution.

[0102] Case 10: Fire detection in high humidity environment Objective: To evaluate the system's robustness and detection capabilities in extremely high humidity conditions and verify smoke detectors' false alarm protection measures when exposed to moisture interference.

[0103] Environmental setting: Sprinklers are used to continuously spray water mist throughout the tunnel, and the humidity is adjusted to above 90%, simulating a fire detection environment under heavy rain or moisture condensation conditions. Moisture may be confused with smoke, increasing the difficulty of detection.

[0104] Fire source arrangement: Ignite oil fire sources to produce thick black smoke accompanied by high temperature, simulating the typical characteristics of fire in an environment with extremely high humidity. The fire source is set in the lower level of the tunnel.

[0105] Detector placement: Install special smoke detectors with moisture resistance, and place thermal imaging monitoring equipment in key locations to prevent humidity from affecting flame identification.

[0106] Testing focus: Verify whether the smoke sensing module can effectively distinguish smoke from moisture, analyze the system false alarm rate and its countermeasures. Observe the accuracy of the image sensing module under high humidity conditions to ensure that thermal imaging can work normally under water vapor interference.

[0107] Environmental parameters: Set different humidity levels, gradually increasing from 70% to 95%, and record changes in system performance, optimize detector and algorithm configurations, and improve the accuracy of fire perception.

[0108] Condition 11: Combined effect of multiple fire points and thermal flow Objective: To test the system's comprehensive detection capabilities when multiple fires occur simultaneously and are accompanied by changes in thermal airflow, and to evaluate the fire perception system's adaptability to complex environments.

[0109] Fire source arrangement: Ignite fire sources in multiple areas of the upper and lower tunnels, and set up hot air flow guide devices to change the direction and speed of hot air flow to simulate the scene of multiple fire outbreaks and mergers. The fire source material is a mixture of flammable plastics and oils to produce high temperature and thick smoke.

[0110] Detector layout: Multiple groups of temperature sensors and smoke detectors are arranged along the fire source path to record the impact of thermal airflow on the detection results. Image monitoring systems are distributed in the area where the thermal airflow passes to capture flame changes in real time.

[0111] Detection focus: The system needs to analyze the changes in temperature and smoke concentration when multiple fire sources merge in real time, and detect the fire characteristics under the disturbance of thermal airflow. Verify whether the multi-parameter fusion algorithm can effectively identify the merged fire source, prevent false alarms and missed alarms, and ensure the system's stable response to complex fire paths.

[0112] Environmental parameters: Set different thermal flow intensities, from low-speed thermal flow to high-speed disturbance, to verify the adaptability of the fire perception system and the reliability of real-time data fusion. Analyze the system's responsiveness and stability under dynamic airflow changes.

[0113] Through these diverse and in-depth working condition designs, the actual application performance and accuracy of the fire sensing system in a double-layer cable tunnel can be effectively tested. These experimental steps are aimed at optimizing system parameters, improving detection accuracy, and enhancing its robustness in complex environments.

[0114] (2) Fire data collection and analysis Multi-parameter data logging: Real-time monitoring: Distributed fiber temperature sensors (DTS) are deployed in double-layer cable tunnels to achieve high-precision temperature monitoring. Through the fiber optic network, the system can obtain the temperature distribution along the entire length of the tunnel, and use the anti-interference characteristics of optical fiber to ensure stable operation in high humidity or dusty environments. High-sensitivity laser smoke detectors use dynamic filtering algorithms to quickly detect fine smoke particles. Smoke detectors integrate adaptive filters based on neural networks, which can distinguish fire smoke from non-fire factors (such as dust and water vapor) in real time to improve detection accuracy. Deploy thermal imaging and visible light cameras. Thermal imaging cameras use infrared thermal imaging algorithms to monitor heat anomalies, which are particularly suitable for operation in low-light environments; visible light cameras are equipped with deep learning models (such as YOLOv5) to identify flames, smoke and other fire-related features in real time, and combine data with other sensor information for analysis.

[0115] Synchronous data collection: Use high-precision time synchronization protocols (such as IEEE 1588 PTP) to ensure accurate alignment of data from all sensors, establish a unified timestamp, and ensure continuous recording of various types of data at different stages of fire development (initial, spread, peak, and weakening) to form a multi-dimensional fire data set.

[0116] Data comparison and analysis: Experimental data comparison: Error analysis tools (such as root mean square error RMSE) are used to compare the data collected by the system with the actual fire development, analyze the detection accuracy of the system, especially in the process of fire spreading from upper to lower floors, to ensure the accuracy and real-time nature of the data. Analyze the changing patterns of temperature, smoke and image data when a fire occurs, and identify potential sources of missed reports and false alarms.

[0117] Multiple rounds of experimental verification: Repeated experiments are conducted and statistical methods (such as t-test and chi-square test) are used to evaluate the false alarm rate, missed alarm rate and response speed of the system under different fire scenarios and environmental conditions to ensure that the results are statistically significant. Data analysis will help identify the weak links of the system under specific conditions and further optimize the detection algorithm and sensor settings.

[0118] (3) System performance evaluation and optimization Performance indicators evaluation: Sensitivity and response time: Quantify the system's response time to various fire types (such as concealed combustion, rapid spread, local high temperature, etc.), use a timer to accurately measure the time from the system detecting the fire signal to the alarm, evaluate its detection sensitivity in the initial fire and fire spread stage, and pay special attention to the impact of the fire characteristics of the upper and lower layers of the double-layer tunnel on the system performance. Use simulated fire models (such as point source fire model and surface source fire model) to predict the heat distribution and smoke diffusion of the fire at different stages, and analyze the changes in response time in combination with real-time monitoring data.

[0119] Stability test: Long-term stability test is carried out under various environmental interferences (such as strong wind, humidity changes, dust, etc.), recording the data consistency and anti-interference ability of the system under different conditions to ensure the reliability and effectiveness of the fire sensing system in practical applications. The system needs to verify its performance stability in long-term operation, especially under extreme conditions.

[0120] Optimization and improvement: Detection algorithm optimization: Based on experimental data, the long short-term memory network (LSTM) model in the temperature sensing module is adjusted, and the learning rate and memory window are optimized to improve the learning ability of fire characteristics. The smoke detection module uses the Kalman filter algorithm to smooth the signal and enhance resistance to environmental interference. The convolutional neural network (CNN) used in the image recognition module can improve the efficiency and accuracy of feature extraction through transfer learning (such as fine-tuning with pre-trained models). Genetic algorithms and particle swarm optimization algorithms are used to optimize the threshold settings of detectors to improve the system's sensitivity and accuracy to fire.

[0121] Sensor layout adjustment: Based on the experimental results, optimize the installation position of the sensor, pay special attention to the air flow and heat distribution of the upper and lower layers of the double-layer tunnel, and ensure that the detectors cover all areas without blind spots. Use fluid dynamics simulation software (such as ANSYS Fluent) to simulate the airflow and temperature distribution to determine the best sensor layout plan to ensure that relevant data can be captured in a timely and accurate manner when a fire occurs.

[0122] The present invention aims at the fire perception and experimental verification method of double-layer cable tunnels, and forms a unique and efficient fire monitoring system from three aspects: the structural characteristics of double-layer tunnels, the multi-parameter verification technology of fire perception, and on-site fire experimental verification, which has remarkable application effects.

[0123] (1) Response to the structural characteristics of double-layer tunnels: The special structure of double-layer tunnels makes the fire sensing method unique. The complex airflow dynamics, temperature distribution and smoke diffusion pattern inside the tunnel make it difficult to effectively apply the traditional single fire detection technology. This invention conducts an in-depth analysis of the air flow characteristics of the upper and lower layers of the tunnel and designs a targeted detector layout scheme to ensure that the temperature and smoke sensors can effectively cover every key area of ​​the tunnel. By comprehensively considering the fire spread path of the double-layer structure, the detection strategy is optimized to ensure the accurate capture and identification of early fire signals.

[0124] (2) Multi-parameter verification improves fire perception accuracy: The fire perception system uses multi-parameter verification technology, combined with temperature, smoke and image data, to significantly improve detection accuracy. The system uses distributed fiber temperature sensors (DTS) to accurately monitor temperature changes, and responds to potential fire signals in a timely manner by setting dynamic temperature thresholds and intelligent temperature analysis algorithms. At the same time, laser smoke detectors monitor smoke concentrations with high sensitivity, and optimize the layout of detectors through smoke diffusion models to reduce the impact of environmental interference. The image perception module uses thermal imaging and visible light image analysis, combined with deep learning algorithms (such as convolutional neural networks (CNNs)), to monitor flame morphology and smoke characteristics in real time, achieving effective fusion of multi-dimensional signals to ensure the accuracy and timeliness of fire identification.

[0125] (3) On-site fire experiment to verify the effectiveness of fire perception: To ensure the effectiveness of the fire perception system, the present invention has designed a rigorous on-site fire experiment verification method. By constructing a double-layer cable tunnel structure that meets the actual engineering requirements at the experimental base, a variety of fuels (such as cables, wood and oil) are set to simulate different types of fire scenes. During the experiment, various sensor data are monitored and recorded in real time, and data is collected synchronously to ensure accurate alignment of temperature, smoke and image information. Through multiple rounds of experiments, the system detection data is compared with the actual fire development situation, and the system's detection accuracy, false alarm rate and response speed are analyzed to provide a reliable basis for subsequent system optimization. In addition, the performance index evaluation will cover key indicators such as sensitivity, response time and stability to ensure the efficient and stable performance of the system in practical applications.

[0126] In summary, the fire perception and experimental verification method for double-layer cable tunnels in the present invention forms an efficient, accurate and reliable fire monitoring solution by combining unique structural characteristics, multi-parameter detection and field verification technology. This solution can not only significantly reduce the false alarm rate of fires and enhance the emergency response capability of tunnel fires, but also has broad application prospects.

[0127] In yet another embodiment of the present invention, a fire sensing system for a double-layer cable tunnel structure is provided, which can be used to implement the fire sensing method for the double-layer cable tunnel structure. Specifically, the system includes: The data acquisition module is used to collect temperature, smoke and image data of the double-layer cable tunnel respectively and to link the double-layer data; A model building module is used to build a temperature distribution model, a smoke diffusion model and an image target detection model according to the collected temperature, smoke and image data; The fusion discrimination module is used to fuse multimodal data of temperature, smoke and image based on the temperature distribution model, smoke diffusion model and image target detection model. When the fused data meets the set fire threshold, it is determined that a fire has occurred.

[0128] The division of modules in the embodiments of the present invention is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present invention may be integrated into one processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0129] In another embodiment of the present invention, a computer device is provided, the computer device comprising a processor and a memory, the memory being used to store a computer program, the computer program comprising program instructions, and the processor being used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of the fire perception method of the double-layer cable tunnel structure.

[0130] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the fire perception method for the double-layer cable tunnel structure in the above embodiment.

[0131] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0132] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0133] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0134] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A fire sensing method for a double-layer cable tunnel structure, characterized in that: include: Collect temperature, smoke and image data of double-layer cable tunnels respectively, and link the double-layer data; According to the collected temperature, smoke and image data, a temperature distribution model, a smoke diffusion model and an image target detection model are constructed respectively; Multimodal data fusion of temperature, smoke and image is performed based on the temperature distribution model, smoke diffusion model and image target detection model. When the fused data meets the set fire threshold, the fire is determined to have occurred.

2. The fire sensing method for a double-layer cable tunnel structure according to claim 1 is characterized in that: The collecting of temperature, smoke and image data of the double-layer cable tunnel includes: Temperature detectors are arranged at the top ceiling of each upper tunnel section, at the intersection nodes, corners and branches of the upper tunnel, and in the area where cables are densely distributed in the upper tunnel. The distributed temperature detectors can monitor the accumulation and conduction of heat in real time. Laser smoke detectors are arranged at each smoke exhaust port, as well as at the entrance and exit of the upper tunnel to monitor the spread of fire smoke. Thermal imaging and visible light camera combination devices are arranged at intervals in each tunnel section. Thermal imaging is used to monitor temperature anomalies, and visible light cameras are used to identify smoke and flames. Laser smoke detectors are arranged in the lower tunnels near the tunnel side walls or at low intervals along both sides of the tunnel; Temperature detectors are arranged at intervals on each cable tray in the lower tunnel and on the equipment boxes to monitor heat accumulation and conduction in real time; visible light cameras are arranged at each tunnel exit and emergency evacuation channel to monitor the evacuation path of personnel; thermal imaging cameras are arranged in the cable tray area to monitor thermal anomalies of the cables in real time.

3. The fire sensing method for a double-layer cable tunnel structure according to claim 2 is characterized in that: The two-layer data linkage includes: All detector signals are transmitted to the monitoring center via wired or wireless networks. The monitoring center adjusts the sensitivity of the detectors in real time based on the detector data. Establish a data linkage mechanism between the two layers: according to any abnormal data in the temperature, smoke and image data of any layer, adjust the sensitivity of the detector on the other layer and enter the high-sensitivity mode.

4. The fire sensing method for a double-layer cable tunnel structure according to claim 1 is characterized in that: The step of constructing a temperature distribution model comprises: Deploy a fiber optic temperature sensor network in the double-layer tunnel, using time division multiplexing (TDM) for data transmission; collect temperature distribution data along the length of the tunnel in real time, sense temperature changes through optical fiber, and use the Raman scattering principle to accurately locate the temperature change point; set multiple temperature alarm thresholds, and when the temperature reaches the threshold, the system triggers different levels of warnings based on the temperature change rate; The long short-term memory network (LSTM) is used to process historical temperature data and analyze the temperature change trend. According to the prediction results output by the LSTM model, the temperature alarm threshold is dynamically adjusted. The support vector machine (SVM) is used to classify the temperature data and distinguish between normal fluctuations and signs of fire.

5. The fire sensing method for a double-layer cable tunnel structure according to claim 1, characterized in that: The smoke diffusion model is constructed, comprising: The detector uses the principle of laser scattering to detect fine smoke particles, and uses an optical sensor array to improve detection sensitivity; an adaptive filtering algorithm is used to eliminate dust and water vapor interference from non-fire smoke; and by analyzing the rate of change of smoke concentration, it is determined whether the fire is fast-spreading or slow-developing. Combined with the Navier-Stokes equations of fluid mechanics, a smoke diffusion model is established. The installation position of the smoke detector is optimized according to the model results. The smoke concentration data collected by the sensor is compared with the model prediction value in real time, and the least squares method is used to correct the abnormal smoke concentration detection results.

6. The fire sensing method for a double-layer cable tunnel structure according to claim 1, characterized in that: The step of constructing an image target detection model comprises: Data is transmitted via the serial port or Ethernet, and convolutional neural network (CNN) is used for image processing. The model is trained to identify flame morphology and smoke characteristics. The model is trained using an image dataset containing various fire scenes. The YOLO real-time target detection algorithm is combined to identify smoke and open flames.

7. The fire sensing method for a double-layer cable tunnel structure according to claim 1, characterized in that: The method of performing multimodal data fusion of temperature, smoke and image based on the temperature distribution model, smoke diffusion model and image target detection model, and determining the occurrence of fire when the fused data meets the set fire threshold, includes: The Bayesian network algorithm is used to fuse image data with temperature and smoke signals. The confidence level of each parameter is calculated and the threshold for fire determination is set. When all signals meet the fire characteristics at the same time, a fire is determined to have occurred. The random forest model is used to train historical fire data, and the sensitivity parameters of the detector are automatically corrected according to the real-time monitoring data. When a sensor sends a fire signal, multiple parameters are called for verification, and the temperature, smoke and image features are integrated. The decision tree algorithm is used to check the data consistency. If all signals are consistent, a fire alarm is triggered. If the signals are inconsistent, a deep analysis process is entered, and anomaly detection is performed through a clustering algorithm.

8. Fire sensing system for double-layer cable tunnel structure, characterized in that: include: The data acquisition module is used to collect temperature, smoke and image data of the double-layer cable tunnel respectively and to link the double-layer data; A model building module is used to build a temperature distribution model, a smoke diffusion model and an image target detection model according to the collected temperature, smoke and image data; The fusion discrimination module is used to fuse multimodal data of temperature, smoke and image based on the temperature distribution model, smoke diffusion model and image target detection model. When the fused data meets the set fire threshold, it is determined that a fire has occurred.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the fire sensing method for a double-layer cable tunnel structure as claimed in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the fire sensing method for a double-layer cable tunnel structure as claimed in any one of claims 1 to 7 are implemented.

Images