Intelligent monitoring operation method and system for tunnel and road fire fighting

Abstract

The invention discloses an intelligent monitoring operation method and system for tunnel road fire fighting, and relates to the technical field of tunnel fire fighting safety, and the method comprises the steps: inputting a vibration feature vector into a pre-trained fire classification model to obtain a fire event, and calculating the position of a fire source through time domain reflection positioning; a deep reinforcement learning model is constructed based on the fire source position and the tunnel state parameters, and a fire-fighting robot path and a water pump start-stop instruction are generated; controlling the fire-fighting robot to move according to the fire-fighting robot path and the water pump start-stop instruction, adjusting water pump injection parameters, and collecting fire diffusion data in real time to generate a feedback signal; updating parameters of the reinforcement learning model according to the feedback signal, dynamically optimizing a fire-fighting robot path and a water pump start-stop instruction, and re-executing; accurate vibration feature vectors are extracted through wavelet noise reduction processing, high-sensitivity monitoring and early warning of fire hazards are achieved, and the intelligent level and emergency response capacity of tunnel and road fire prevention and control are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel fire safety, and in particular to an intelligent monitoring and operation method and system for tunnel road fire fighting. Background Art

[0002] In recent years, with the continuous advancement of fiber optic sensing technology, artificial intelligence algorithms, and automated control systems, tunnel fire monitoring systems have gradually developed towards intelligent, multimodal sensing. Existing research has used distributed optical fiber to collect temperature or vibration signals, combined with image recognition and smoke detection to achieve early warning of fires. These systems also utilize pre-set control logic to drive firefighting equipment for emergency response. This has improved the accuracy and response speed of tunnel fire monitoring to a certain extent. However, due to the narrow and long interior spaces and limited ventilation conditions of tunnels, fires can spread rapidly, posing significant challenges to evacuation and firefighting.

[0003] Tunnel fire monitoring methods mainly focus on monitoring parameters such as temperature and smoke concentration. Such methods often fail to provide sufficient early warning information, especially in the early stages of a fire, and are prone to false alarms or missed alarms, resulting in inefficient dispatch of firefighting resources. Summary of the Invention

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

[0005] Therefore, the present invention provides an intelligent monitoring and operation method for tunnel road fire fighting to solve the problem of low fire source identification accuracy.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides an intelligent monitoring and operation method for tunnel highway fire fighting, which includes collecting vibration signals through distributed optical fibers on the side walls of the tunnel, and performing wavelet noise reduction processing to extract vibration features to obtain vibration feature vectors; inputting the vibration feature vectors into a pre-trained fire classification model to obtain fire events, and calculating the fire source location through time domain reflection positioning; constructing a deep reinforcement learning model based on the fire source location and tunnel state parameters to generate a fire robot path and water pump start-stop instructions; controlling the fire robot to move according to the fire robot path and water pump start-stop instructions and adjust the water pump injection parameters, and collecting fire spread data in real time to generate a feedback signal; updating the reinforcement learning model parameters according to the feedback signal, dynamically optimizing the fire robot path and water pump start-stop instructions and re-executing them; setting a fire safety threshold according to the tunnel cross-sectional area, and evaluating the fire status through the fire safety threshold to detect the fire extinguishing situation.

[0007] As a preferred solution of the intelligent monitoring and operation method for tunnel road fire fighting according to the present invention, the vibration signal is collected through distributed optical fibers on the tunnel sidewall, and wavelet noise reduction processing is performed to extract vibration characteristics to obtain vibration feature vectors. The specific steps are as follows: The distributed optical fiber on the tunnel sidewall collects the tunnel structure health status, external influences and traffic loads to form a vibration signal; The noise and interference signals in the vibration signal are removed by filtering operation, the vibration signal is analyzed by wavelet transform, the vibration signal is decomposed into different frequency components, and the vibration signal is reconstructed by inverse wavelet transform to obtain the vibration signal after noise reduction; The vibration signal after noise reduction is analyzed by wavelet transform, and the features in time domain, frequency domain and wavelet energy distribution are extracted. The extracted multiple features are combined into a feature vector to obtain the vibration feature vector.

[0008] As a preferred solution of the intelligent monitoring and operation method for tunnel road fire fighting described in the present invention, the vibration feature vector is input into a pre-trained fire classification model to obtain a fire event, and the fire source location is calculated by time domain reflectometry positioning. The specific steps are as follows: The vibration feature vector is input into the pre-trained fire classification model for analysis to obtain the fire distribution probability and output the category label; According to the category label, the time-domain reflection positioning function of the distributed optical fiber is triggered. By sending a pulse signal to the optical fiber and receiving the reflected signal, the specific position of the fire source on the optical fiber is calculated based on the signal propagation time and reflection characteristics. Combined with the layout path of the optical fiber on the tunnel side wall, the fire source position is mapped to the actual tunnel space coordinates.

[0009] As a preferred solution of the intelligent monitoring and operation method for tunnel road fire fighting described in the present invention, a deep reinforcement learning model is constructed based on the fire source location and tunnel state parameters to generate the fire robot path and water pump start and stop instructions. The specific steps are as follows: The state space is formed based on the fire source location, tunnel state parameters, and the state of the firefighting equipment. The action space is defined based on the range of the firefighting robot and the on / off action of the water pump. By combining deep reinforcement learning algorithms with neural networks, a deep reinforcement learning model is constructed based on state space and action space; By combining simulation software with the actual tunnel environment, a virtual environment for tunnel fire simulation is constructed. In this virtual environment, a deep reinforcement learning model is optimized through a continuous interactive state-action-reward cycle. The state space is input into the optimized deep reinforcement learning model to generate the movement path of the fire-fighting robot and the start and stop control commands of the water pump.

[0010] As a preferred solution of the intelligent monitoring and operation method for tunnel road fire fighting described in the present invention, the fire fighting robot is controlled to move according to the fire fighting robot path and water pump start and stop instructions and to adjust the water pump spray parameters, and to collect fire spread feedback signals in real time. The specific steps are as follows: The firefighting robot moves from its current position toward the fire source according to the planned path. When the firefighting robot approaches the fire source, it starts the water pump and adjusts the spray parameters according to the start and stop control commands of the water pump. Through distributed optical fibers, fire spread data is collected and the degree of fire reduction and fire extinguishing efficiency during the fire extinguishing process are recorded. Analysis is then performed to evaluate the effectiveness of the fire extinguishing strategy. The evaluation results are obtained and combined with the operating status of the fire-fighting robot and the working status of the water pump as feedback signals.

[0011] As a preferred solution of the intelligent monitoring and operation method for tunnel road fire fighting according to the present invention, wherein: the reinforcement learning model parameters are updated according to the feedback signal, the fire fighting robot path and the water pump start and stop instructions are dynamically optimized and re-executed, the specific steps are as follows: Input the feedback signal into the deep reinforcement learning model, use the feedback signal to perform online learning on the deep reinforcement learning model, and adjust the parameters of the deep reinforcement learning model through the policy gradient method; Based on the updated deep reinforcement learning model, the optimized firefighting robot's movement path and the water pump's start-stop control commands are recalculated. The firefighting robot readjusts its route according to the optimized firefighting robot path, and readjusts the spray parameters according to the optimized water pump's start-stop control commands after approaching the fire source.

[0012] As a preferred solution of the intelligent monitoring and operation method for tunnel road fire fighting according to the present invention, the fire safety threshold is obtained according to the tunnel cross-sectional area, the fire status is evaluated by the fire safety threshold, and a multi-spectral scanning is started to detect the fire extinguishing status. The specific steps are as follows: Based on the geometric structure of the tunnel, the air circulation capacity and heat diffusion capacity are calculated. The fire safety threshold is obtained by combining the air circulation capacity and heat diffusion capacity according to the tunnel parameters, fire dynamics theory and tunnel cross-sectional area. Use distributed optical fiber to monitor temperature distribution, smoke concentration, and flame coverage in the tunnel, and compare the collected data with fire safety thresholds to assess the fire status; Based on the fire status, a multi-spectral scan is initiated to conduct a comprehensive scan of the fire area to detect the fire extinguishing status.

[0013] In a second aspect, the present invention provides an intelligent monitoring and operation system for tunnel road fire fighting, comprising an acquisition module, a positioning module, an instruction generation module, an execution module, an optimization module, and a detection module; The acquisition module is used to collect vibration signals through distributed optical fibers on the side walls of the tunnel, and perform wavelet noise reduction processing to extract vibration features to obtain vibration feature vectors; the positioning module is used to input the vibration feature vectors into a pre-trained fire classification model to obtain fire events, and calculate the fire source location through time domain reflection positioning; the instruction generation module is used to construct a deep reinforcement learning model based on the fire source location and tunnel state parameters, and generate the fire robot path and water pump start-stop instructions; the execution module is used to control the fire robot to move according to the fire robot path and water pump start-stop instructions and adjust the water pump injection parameters, and collect fire spread data in real time to generate feedback signals; the optimization module is used to update the reinforcement learning model parameters according to the feedback signal, dynamically optimize the fire robot path and water pump start-stop instructions and re-execute them; the detection module is used to set the fire safety threshold according to the tunnel cross-sectional area, evaluate the fire status through the fire safety threshold, and detect the fire extinguishing situation.

[0014] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the intelligent monitoring and operation method for tunnel road fire fighting as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent monitoring and operation method for tunnel road fire fighting as described in the first aspect of the present invention.

[0016] The present invention achieves the following beneficial effects: by collecting vibration signals through distributed optical fibers on the tunnel sidewalls and extracting precise vibration eigenvectors using wavelet noise reduction, it achieves highly sensitive monitoring and early warning of fire hazards, effectively reducing false alarm rates and improving fire source location accuracy. Furthermore, time-domain reflectometry technology is used to precisely locate the fire source, providing critical information and significantly improving the intelligent level of tunnel and highway fire prevention and control and emergency response capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 The present invention is a flow chart of an intelligent monitoring and operation method for fire fighting in tunnel highways.

[0019] Figure 2 Schematic diagram of the intelligent monitoring and operation system for tunnel highway fire fighting.

[0020] Figure 3 Flowchart for extracting vibration features for wavelet denoising.

[0021] Figure 4 A flowchart for building and optimizing deep reinforcement learning models. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0023] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

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

[0025] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides an intelligent monitoring and operation method for tunnel road fire fighting, comprising the following steps: S1. Vibration signals are collected through distributed optical fibers on the tunnel sidewalls, and wavelet noise reduction is performed to extract vibration features and obtain vibration feature vectors.

[0026] Distributed optical fibers on the tunnel sidewalls collect information about the tunnel's structural health, external influences, and traffic loads to generate vibration signals.

[0027] Specifically, distributed optical fiber is deployed on the side walls of the tunnel. The distributed optical fiber can monitor in real time tiny vibrations caused by structural changes, external environmental factors (such as earthquakes and ground movement), and passing vehicles. The optical signal transmitted in the optical fiber is analyzed using optical time domain reflectometry (OTDR) technology to detect changes in the optical signal inside the optical fiber caused by the health of the tunnel structure, external influences, and traffic load factors, forming a vibration signal reflecting the status of the tunnel.

[0028] The noise and interference signals in the vibration signal are removed by filtering operation, the vibration signal is analyzed by wavelet transform, the vibration signal is decomposed into different frequency components, and the vibration signal is reconstructed by inverse wavelet transform to obtain the noise-reduced vibration signal.

[0029] Specifically, a filter (such as a bandpass or low-pass filter) is applied to remove obvious noise and interference signals, and the filtered vibration signal is decomposed by wavelet transform to decompose the vibration signal into wavelet coefficients of different frequency bands. After completing the wavelet decomposition, an adaptive denoising threshold is obtained based on an analysis method that combines the local singularity of the signal with the statistical characteristics of the noise. The high-frequency detail coefficients are processed using the adaptive denoising threshold, and the high-frequency wavelet coefficients whose absolute values are less than the adaptive denoising threshold are set to zero. The processed low-frequency approximation coefficients and high-frequency detail coefficients are reconstructed through inverse wavelet transform. At the same time, the high-frequency wavelet coefficients greater than or equal to the adaptive denoising threshold are shrunk to retain structural information. The adjusted wavelet coefficients are reconstructed through inverse wavelet transform to obtain the denoised vibration signal.

[0030] It should be noted that based on the characteristics of the vibration signal (such as signal-to-noise ratio and frequency distribution), statistical analysis methods are used to estimate the noise level. For example, the noise intensity is assessed by calculating the standard deviation of the high-frequency coefficients after wavelet decomposition. The threshold is set using the universal threshold method based on the noise intensity.

[0031] The vibration signal after noise reduction is analyzed by wavelet transform, and the features in time domain, frequency domain and wavelet energy distribution are extracted. The extracted multiple features are combined into a feature vector to obtain the vibration feature vector.

[0032] Specifically, wavelet transform is applied to perform multi-level decomposition to obtain details and approximation coefficients in different frequency bands, and to analyze signal characteristics in detail in the time and frequency domains. Various features are extracted from the decomposition levels: statistical parameters such as the mean, variance, and peak value of the vibration signal are calculated in the time domain; and characteristics such as the energy distribution and main frequency components of each frequency band are analyzed in the frequency domain. Based on the wavelet energy distribution, the wavelet energy percentage of each decomposition level is calculated to reflect the energy distribution of the vibration signal at different scales. Multiple features extracted from the time domain, frequency domain and wavelet energy distribution are combined into a vibration feature vector.

[0033] S2. Input the vibration feature vector into the pre-trained fire classification model to obtain the fire event, and calculate the fire source location through time domain reflectometry positioning.

[0034] The vibration feature vector is input into the pre-trained fire classification model for analysis to obtain the fire distribution probability and output the category label.

[0035] Specifically, a large amount of vibration signal data, including normal conditions and different types of fire conditions, is collected and labeled. Filtering, denoising, and feature extraction are performed to form vibration feature vectors. An appropriate deep learning model architecture is selected, with the input defined as the vibration feature vector and the output as the probability distribution of fire categories. Cross entropy is set as the loss function, and optimization algorithms such as Adam or SGD are used to update the parameters. Divide the collected fire-related vibration signal data into training and test sets. Perform multiple rounds of iterative training using the training set data. In each round, calculate the error between the predicted value and the true label through forward propagation, and then adjust the weights through backpropagation. Use the validation set to monitor the performance of the pre-trained fire classification model, and apply regularization techniques or early stopping methods as appropriate to prevent overfitting. After completing all training rounds, use an independent test set to evaluate the overall performance of the pre-trained fire classification model. Save the trained fire classification model and weights. Preprocess the extracted vibration feature vectors and reshape them into a two-dimensional array (using Python and TensorFlow frameworks). For example, normalize or standardize the vibration feature vectors to ensure they fall within the appropriate range. Convert them into the tensor format required by the fire classification model (such as a NumPy array or a PyTorch / TensorFlow tensor). The processed vibration feature vector is input into the trained fire classification model. The pre-trained fire classification model uses a multi-layer neural network (such as fully connected layers and convolutional layers) to perform nonlinear transformations and feature mapping on the vibration feature vector. The model then calculates the probability distribution of each fire category. The softmax function is used to convert the raw scores output by the fire classification model into probability values, representing the likelihood that the vibration signal corresponds to each fire category (e.g., no fire, open flame, smoldering fire). Based on the probability distribution of fire categories, the category label corresponding to the highest probability is selected as the output.

[0036] According to the category label, the time-domain reflection positioning function of the distributed optical fiber is triggered. By sending a pulse signal to the optical fiber and receiving the reflected signal, the specific position of the fire source on the optical fiber is calculated based on the signal propagation time and reflection characteristics. Combined with the layout path of the optical fiber on the tunnel side wall, the fire source position is mapped to the actual tunnel space coordinates.

[0037] Specifically, based on the category label, if the classification result is a fire event, the distributed optical fiber time domain reflectometry positioning function is immediately triggered. A short pulse of light signal is sent to the distributed optical fiber laid along the tunnel sidewall. When the short pulse of light encounters uneven points in the fiber (such as areas with temperature fluctuations caused by fire), it will be reflected. The reflected light signals are then received and the arrival time of each reflected light signal is recorded. Based on the speed of light in an optical fiber (approximately two-thirds the speed of light in a vacuum), the distance from the reflection point to the fiber's starting point is calculated. By analyzing the time differences and intensity variations of multiple reflected signals, the coordinates of the most likely fire source on the optical fiber are determined. Combined with the specific layout of the optical fiber along the tunnel sidewall (for example, a detailed layout of the optical fiber along the tunnel roof or walls), the relative position mapping on the optical fiber is converted into actual tunnel spatial coordinates, thereby accurately locating the actual three-dimensional position of the fire source.

[0038] The expression for calculating the distance from the fire source should be explained as follows:

[0039] Among them, L is the distance from the fire source, v is the speed of light, and t is the round-trip time of the reflected light signal.

[0040] S3. Build a deep reinforcement learning model based on the fire source location and tunnel state parameters to generate the firefighting robot path and water pump start and stop instructions.

[0041] The state space is composed of the actual three-dimensional position of the fire source, the tunnel state parameters and the state of the fire-fighting equipment. The action space is defined according to the activity range of the fire-fighting robot and the switching action of the water pump.

[0042] Specifically, temperature distribution and smoke concentration data are collected in real time in the tunnel through distributed optical fibers, infrared thermal imagers, laser scattering smoke sensors, high-definition cameras and other equipment, and the raw data are cleaned using algorithms such as Kalman filtering and mean filtering to eliminate noise; a GPS / inertial navigation module is installed on the fire-fighting robot to obtain real-time position, the battery management system is connected to read power information, the remaining amount of fire extinguishing agent is monitored through a liquid level sensor, and the on / off status is obtained by connecting to the water pump control circuit. The position coordinates are standardized according to the tunnel mileage, the power and water level are converted into percentages, and the water pump status is encoded as 0 / 1 values to obtain a standardized numerical representation of the equipment information. The temperature distribution and smoke concentration data of the monitoring points in the tunnel are selected and the data is cleaned; the equipment information such as the current position of the fire-fighting robot, battery power, remaining amount of fire extinguishing agent, and on / off status of the water pump are converted into standardized numerical representations to obtain tunnel status parameters and the status of the fire-fighting equipment; The actual three-dimensional location and intensity of the fire source are determined by using distributed optical fibers, a heat release rate formula based on the properties of the burning material and mass flow rate, and multi-source data fusion algorithms such as Kalman filtering. The fire source's coordinates, spatial distribution characteristics, and heat release parameters (HRR, temperature gradient, smoke generation rate, etc.) in the tunnel are converted into precise numerical values. Real-time calibration is used to improve accuracy and normalization is performed to fit within a uniform numerical range. These values are combined into a high-dimensional vector in a sequential order (e.g., fire source information, then environmental parameters, and finally equipment status) to form a complete state space representation. Each firefighting robot position is assigned a unique identifier. For each position, the robot's executable action options are listed, including basic movement commands such as forward, backward, left turn, right turn, and stop. Considering the different operating modes of the water pump, operations (such as starting, stopping, and adjusting water pressure) are encoded as independent action commands and integrated into the same action space with the movement commands. By constructing a joint action set, where each action consists of a location identifier and one or more operation commands, the optimal action strategy can be selected based on the current state.

[0043] By combining deep reinforcement learning algorithms with neural networks, a deep reinforcement learning model is constructed based on state space and action space.

[0044] Specifically, a deep reinforcement learning model (using a multi-layer perceptron convolutional neural network) is used as the neural network input. Training is performed using a large amount of simulated fire scene data. A reward function (e.g., rewards based on firefighting efficiency and the degree of protection provided to tunnel facilities) is employed during training. Deep reinforcement learning algorithms (such as DQN and A3C) are used to continuously optimize the neural network parameters, allowing the model to learn to select the optimal actions to control and extinguish fires under different fire conditions. The deep reinforcement learning model parameters are initialized, and the required hyperparameters for training, such as the learning rate, discount factor, and experience replay buffer size, are set. The current environmental state is then fed into the neural network for processing, and the neural network outputs the value or probability distribution of each possible action under a given state. Through interaction with the environment, a deep reinforcement learning model is developed that can make intelligent decisions based on the specific circumstances of a tunnel fire scene.

[0045] By combining simulation software with the actual environment of the tunnel, a virtual environment for tunnel fire simulation is constructed. In this virtual environment, the deep reinforcement learning model is optimized through continuous interactive state-action-reward cycles.

[0046] Specifically, a 3D geometric model of the tunnel was created in the selected simulation software. The tunnel's length, width, and height were set according to actual dimensions, and material properties (such as thermal conductivity and fire resistance) were defined. The software's physics engine was used to configure the dynamic characteristics of the fire, simulating the flame spread process by setting the fire source location, heat release rate, and type of burning material. Temperature field and smoke diffusion modules were added to simulate the flow of hot air and smoke concentration within the tunnel using computational fluid dynamics (CFD) methods. Distributed optical fiber was integrated to collect real-time temperature and vibration data within the virtual environment. The entire virtual environment was verified and debugged to ensure that fire behavior was consistent with physical laws. Virtual environment parameters were adjusted to improve simulation accuracy, resulting in a virtual environment that dynamically reflects the characteristics of tunnel fires. A pre-defined state space and action space were deployed within the virtual environment, enabling the deep reinforcement learning model to interact with it. Upon launching the simulation, the firefighting robot executed selected actions based on its current state and received real-time feedback from the fire environment, including new state information and corresponding rewards (calculated based on factors such as fire extinguishing efficiency and energy consumption). Data from the state-action-reward loop was collected for subsequent analysis. The deep reinforcement learning model is trained using data from the state-action-reward cycle, and the parameters of the deep reinforcement learning model are adjusted through batch updates or online learning. This process is repeated continuously, so that the deep reinforcement learning model can learn the optimal behavior pattern after a large number of iterations in the virtual environment.

[0047] The state space is input into the optimized deep reinforcement learning model to generate the movement path of the fire-fighting robot and the start and stop control commands of the water pump.

[0048] Specifically, the current state space data (including the fire source location, tunnel environmental parameters, and firefighting equipment status) is input into an optimized deep reinforcement learning model. Based on this input state information, the deep reinforcement learning model uses its internal neural network to calculate the value or probability distribution of each possible action and selects the optimal action as output. For firefighting robots, the output actions include specific movement path points or directional instructions (such as forward or left turn), thereby generating the optimal path from the current location to the target fire source. For water pumps, the output actions include start / stop commands or water pressure adjustment instructions.

[0049] S4. Control the fire-fighting robot to move according to the fire-fighting robot path and water pump start and stop instructions and adjust the water pump injection parameters, and collect fire spread data in real time to generate feedback signals.

[0050] The fire-fighting robot moves from its current position to the fire source according to the planned path. When the fire-fighting robot approaches the fire source, it starts the water pump and adjusts the spray parameters according to the start and stop control commands of the water pump.

[0051] Specifically, the firefighting robot moves from its current position toward the fire source according to the planned path, continuously monitoring the distance between itself and the fire source. When the robot approaches the fire source to within a preset distance threshold (the most effective firefighting distance is between 10 and 20 meters from the fire source, and for safety and operational flexibility, a distance threshold of 15 meters can be selected as the standard for triggering the water pump to start), the water pump start command is triggered; Furthermore, the robot follows a pre-calculated optimal path, utilizing its navigation system (such as LiDAR, GPS, or a visual recognition system) to correct its direction and position in real time, ensuring precise arrival at the target location. As it approaches the fire source, a deep reinforcement learning model uses the latest state-space data to issue a water pump start command and simultaneously determine appropriate spray parameters, such as water pressure and spray angle. Upon receiving the pump start command, the pump control system immediately responds, adjusting to the specified operating parameters to effectively control the fire.

[0052] Through distributed optical fibers, fire spread data is collected and the degree of fire reduction and fire extinguishing efficiency during the fire extinguishing process are recorded. Analysis is then performed to evaluate the effectiveness of the fire extinguishing strategy. The evaluation results are obtained and combined with the operating status of the fire-fighting robot and the working status of the water pump as feedback signals.

[0053] Specifically, distributed optical fibers are used to collect data on the spread of fire in the tunnel in real time, including indicators such as temperature changes and smoke concentration, and to record the degree of fire reduction and fire extinguishing efficiency from the start to the end of fire extinguishing. During the fire extinguishing process, distributed optical fibers continuously monitor and transmit data to the central processor. For a medium-sized tunnel fire scenario, the initial rapid response stage is set to analyze the fire intensity every 30 seconds; as the fire extinguishing work progresses, if the fire is observed to have significantly weakened, the time interval is adjusted to every 5 minutes; before confirming that the fire has been completely extinguished, the final inspection is restored to every 1 minute. By collecting temperature distribution and smoke concentration data from monitoring points in the tunnel every minute, as well as information such as the position, power level, remaining amount of fire extinguishing agent, and water pump status of the fire-fighting robot, the changes in fire intensity are analyzed based on preset time intervals. Standardized evaluation indicators are constructed using parameters such as heat release rate (HRR), temperature gradient, smoke concentration, and flame area. The fire extinguishing effect value is calculated using a weighted formula. Combined with time series trend analysis and judgment (such as HRR drops to 10% of the initial value and the temperature drops to a safe range), the fire extinguishing effect is dynamically evaluated. The effectiveness of fire suppression strategies is quantified by calculating key performance indicators (KPIs), such as the temperature drop rate in the fire area, the smoke density reduction ratio, and the required extinguishing time. Supplementary information is collected, including the operating status of the firefighting robot (e.g., movement speed, path accuracy) and the operating status of the water pump (e.g., number of starts, operating hours, and frequency of water pressure adjustment). This feedback signal is generated by combining the fire suppression effectiveness assessment with this supplementary information.

[0054] It should be noted that the key performance indicators of fire extinguishing effectiveness include the temperature drop rate in the fire source area, the smoke concentration reduction ratio, the required fire extinguishing time, and the amount and efficiency of fire water used. The temperature drop rate in the fire source area reflects the efficiency of fire extinguishing measures in reducing the temperature at the fire scene. The smoke concentration reduction ratio evaluates the effect of fire extinguishing actions on improving air quality. A shorter fire extinguishing time usually means a more efficient fire extinguishing strategy.

[0055] S5. Update the reinforcement learning model parameters based on the feedback signal, dynamically optimize the firefighting robot path and water pump start and stop instructions, and re-execute them.

[0056] The feedback signal is input into the deep reinforcement learning model, the feedback signal is used to perform online learning on the deep reinforcement learning model, and the parameters of the deep reinforcement learning model are adjusted through the policy gradient method.

[0057] Specifically, the feedback signals collected during the firefighting process (including the degree of fire reduction, firefighting efficiency, the operating status of the firefighting robot and the working status of the water pump) are input into the deep reinforcement learning model as new training data. The feedback signals are used to calculate the reward value of the current strategy, and the quality of the actions taken by the deep reinforcement learning model in the current state is evaluated based on the reward value. The policy gradient method (PPO) calculates the gradient of the loss function based on the relationship between reward values and action probability distributions. Gradient ascent is then used to update the parameters of the deep reinforcement learning model to optimize the policy. A batch of data, including state, action, reward, and next state, is sampled from the experience replay buffer. The advantage function for each action is calculated and used to adjust the weight parameters of the neural network. This repeated process enables the deep reinforcement learning model to learn online and gradually improve its decision-making capabilities.

[0058] Based on the updated deep reinforcement learning model, the optimized firefighting robot's movement path and the start-stop control commands of the water pump are recalculated. The firefighting robot readjusts its path according to the optimized firefighting robot path, and readjusts the injection parameters according to the optimized start-stop control commands of the water pump after approaching the fire source.

[0059] Specifically, the current environmental state (including the fire source location, tunnel conditions, and firefighting equipment status) is input into the deep reinforcement learning model to recalculate the optimized firefighting robot's movement path and water pump start and stop control commands. Based on the new parameters and strategies, the deep reinforcement learning model generates a more efficient path to the fire source and precise water pump operation instructions. The firefighting robot receives the optimized path and uses navigation to adjust its route in real time, accurately advancing toward the fire source. When the robot approaches the fire source and reaches a preset distance, it starts or adjusts the water pump according to the optimized water pump control commands, such as adjusting the spray pressure or angle, to extinguish the fire. Throughout this process, the robot continuously monitors changes in the surrounding environment and feeds new data back to the deep reinforcement learning model. When necessary, the path and water pump control strategy can be further optimized, achieving dynamic response and efficient firefighting.

[0060] S6. Obtain a fire safety threshold based on the tunnel cross-sectional area, assess the fire status using the fire safety threshold, and initiate multispectral scanning to detect the fire extinguishing status.

[0061] According to the geometric structure of the tunnel, the air circulation capacity and heat diffusion capacity are calculated. The fire safety threshold is obtained by combining the air circulation capacity and heat diffusion capacity based on the tunnel parameters, fire dynamics theory and tunnel cross-sectional area.

[0062] Specifically, based on the tunnel's geometric structure, including parameters such as tunnel length, cross-sectional area, longitudinal slope, and ventilation shaft spacing, combined with tunnel ventilation design data, the air flow rate through the tunnel per unit time is calculated by measuring the tunnel entrance wind speed, outlet pressure difference, and ventilation equipment power, thereby obtaining the air circulation capacity. Simultaneously, based on fire dynamics theory, combined with the tunnel's cross-sectional area and data from past fire cases, the heat diffusion capacity during a fire is calculated, focusing on analyzing the effects of fire source power, combustion material type, and ventilation conditions on heat diffusion. The safety factor is dynamically adjusted based on the ratio of the tunnel cross-sectional area to the air circulation capacity. When the tunnel cross-sectional area is greater than 150m² or the ventilation capacity is less than 120,000 m³ / h, the safety factor is increased to 1.5. When the tunnel cross-sectional area is ≤100m² and the ventilation capacity is ≥180,000 m³ / h, the safety factor is reduced to 0.8. The fire safety threshold is set at 800°C based on the comprehensive air circulation capacity, heat diffusion capacity and safety factor, combined with the real-time monitoring data of the temperature sensor in the tunnel.

[0063] Use distributed optical fiber to monitor the temperature distribution, smoke concentration, and flame coverage in the tunnel, and compare the collected fire data with the fire safety threshold to assess the fire status; Specifically, a distributed fiber optic network monitors the temperature distribution, smoke concentration, and flame coverage in the tunnel in real time. Distributed optical fibers are laid out along the tunnel, providing a continuous data stream that transmits collected real-time data (such as the temperature rise in a specific area, smoke concentration levels, and the length or area affected by the flames) to data analysis. This real-time monitoring data is then compared with pre-calculated fire safety thresholds, which include indicators such as maximum allowable temperature rise and smoke concentration limits. The current fire status is assessed based on this comparison: if any one or more parameters exceed the safety threshold, the fire has reached a dangerous level and immediate action is required. If all parameters are within the safe range, the fire is considered to be under control.

[0064] Based on the fire status, a multi-spectral scan is initiated to conduct a comprehensive scan of the fire area to detect the fire extinguishing status.

[0065] Specifically, the system determines whether to activate the multispectral scanning device based on the fire status (such as temperature distribution, smoke concentration, and flame coverage) detected by the distributed optical fiber. If the fire approaches or exceeds the safety threshold, the multispectral scanning device is immediately activated. The fire area is scanned in all directions, using multi-band spectral signals (such as visible light, infrared, and near-infrared) to obtain information such as the thermal radiation distribution of the fire area, the location of residual hot spots, and the dissipation of smoke. Through image processing and spectral feature extraction technology, it detects whether there are open flames or high-temperature areas and evaluates the fire extinguishing effect: if residual flames or abnormally high-temperature spots are found, the specific location is marked and it is determined that the fire has not been completely extinguished; if all fire area indicators meet safety standards, it is confirmed that the fire has been effectively controlled.

[0066] This embodiment also provides an intelligent monitoring and operation system for tunnel road fire fighting, including: a collection module, a positioning module, a command generation module, an execution module, an optimization module and a detection module; The acquisition module is used to collect vibration signals through distributed optical fibers on the tunnel sidewall, and perform wavelet noise reduction processing to extract vibration characteristics and obtain vibration feature vectors; The positioning module is used to input the vibration feature vector into the pre-trained fire classification model to obtain the fire event and calculate the fire source location through time domain reflectometry positioning; The command generation module is used to build a deep reinforcement learning model based on the fire source location and tunnel state parameters to generate the fire robot path and water pump start and stop commands; The execution module is used to control the firefighting robot to move according to the firefighting robot path and water pump start and stop instructions, adjust the water pump injection parameters, and collect fire spread data in real time to generate feedback signals; The optimization module is used to update the reinforcement learning model parameters based on the feedback signal, dynamically optimize the firefighting robot path and the water pump start and stop instructions, and re-execute them; The detection module is used to set a fire safety threshold according to the tunnel cross-sectional area, evaluate the fire status through the fire safety threshold, and detect the fire extinguishing situation.

[0067] This embodiment also provides a computer device for use in an intelligent monitoring and operation method for tunnel road fire fighting, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent monitoring and operation method for tunnel road fire fighting as proposed in the above embodiment.

[0068] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0069] This embodiment also provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the program implements the intelligent monitoring and operation method for tunnel road fire fighting proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0070] In summary, this invention achieves highly sensitive monitoring and early warning of fire hazards by collecting vibration signals through distributed optical fibers on the tunnel sidewalls and extracting precise vibration eigenvectors using wavelet noise reduction. This effectively reduces false alarm rates and improves fire source location accuracy. Furthermore, time-domain reflectometry technology is used to precisely locate the fire source, providing critical information and significantly improving the intelligent level of fire prevention and control in tunnel highways and emergency response capabilities.

[0071] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent monitoring and operation method for tunnel road fire fighting, characterized by: include, Vibration signals are collected through distributed optical fibers on the tunnel sidewalls, and wavelet noise reduction is performed to extract vibration features and obtain vibration feature vectors. The vibration feature vector is input into the pre-trained fire classification model to obtain the fire event, and the fire source location is calculated through time domain reflectometry positioning; A deep reinforcement learning model is built based on the fire source location and tunnel state parameters to generate the fire robot path and water pump start and stop instructions; Control the firefighting robot to move according to the firefighting robot path and water pump start and stop instructions, adjust the water pump injection parameters, and collect fire spread data in real time to generate feedback signals; Update the reinforcement learning model parameters based on the feedback signal, dynamically optimize the firefighting robot path and the water pump start and stop instructions, and re-execute them; A fire safety threshold is set according to the tunnel cross-sectional area, and the fire status is evaluated by the fire safety threshold to detect the fire extinguishing situation.

2. The intelligent monitoring and operation method for tunnel road fire fighting according to claim 1, characterized in that: The vibration signal is collected through the distributed optical fiber on the side wall of the tunnel, and the vibration characteristics are extracted by wavelet noise reduction processing to obtain the vibration feature vector. The specific steps are as follows: The distributed optical fiber on the tunnel sidewall collects the tunnel structure health status, external influences and traffic loads to form vibration signals; The noise and interference signals in the vibration signal are removed by filtering operation, the vibration signal is analyzed by wavelet transform, the vibration signal is decomposed into different frequency components, and the vibration signal is reconstructed by inverse wavelet transform to obtain the vibration signal after noise reduction; The vibration signal after noise reduction is analyzed by wavelet transform, and the features in time domain, frequency domain and wavelet energy distribution are extracted. The extracted multiple features are combined into a feature vector to obtain the vibration feature vector.

3. The intelligent monitoring and operation method for tunnel road fire fighting according to claim 2, characterized in that: The vibration feature vector is input into the pre-trained fire classification model to obtain the fire event, and the fire source position is calculated by time domain reflectometry positioning. The specific steps are as follows: The vibration feature vector is input into the pre-trained fire classification model for analysis to obtain the fire distribution probability and output the category label; According to the category label, the time-domain reflection positioning function of the distributed optical fiber is triggered. By sending a pulse signal to the distributed optical fiber and receiving the reflected signal, the specific position of the fire source on the optical fiber is calculated based on the signal propagation time and reflection characteristics. Combined with the layout path of the distributed optical fiber on the tunnel side wall, the fire source position is mapped to the actual tunnel space coordinates.

4. The intelligent monitoring and operation method for tunnel road fire fighting according to claim 3, characterized in that: The deep reinforcement learning model is constructed based on the fire source location and tunnel state parameters to generate the fire robot path and water pump start and stop instructions. The specific steps are as follows: The state space is formed based on the fire source location, tunnel state parameters, and the state of the firefighting equipment. The action space is defined based on the range of the firefighting robot and the on / off action of the water pump. By combining deep reinforcement learning algorithms with neural networks, a deep reinforcement learning model is constructed based on state space and action space; By combining simulation software with the actual tunnel environment, a virtual environment for tunnel fire simulation is constructed. In this virtual environment, a deep reinforcement learning model is optimized through a continuous interactive state-action-reward cycle. The state space is input into the optimized deep reinforcement learning model to generate the movement path of the fire-fighting robot and the start and stop control commands of the water pump.

5. The intelligent monitoring and operation method for tunnel road fire fighting according to claim 4, characterized in that: The firefighting robot is controlled to move according to the firefighting robot path and the water pump start and stop instructions and adjust the water pump spray parameters, and real-time fire spread data is collected to generate feedback signals. The specific steps are as follows: The firefighting robot moves from its current position toward the fire source according to the planned path. When the firefighting robot approaches the fire source, it starts the water pump and adjusts the water pump spray parameters according to the water pump start and stop control commands. Through distributed optical fibers, fire spread data is collected and the degree of fire reduction and fire extinguishing efficiency during the fire extinguishing process are recorded. Analysis is then performed to evaluate the effectiveness of the fire extinguishing strategy. The evaluation results are obtained and combined with the operating status of the fire-fighting robot and the working status of the water pump as feedback signals.

6. The intelligent monitoring and operation method for tunnel road fire fighting according to claim 5, characterized in that: The reinforcement learning model parameters are updated according to the feedback signal, the fire fighting robot path and the water pump start and stop instructions are dynamically optimized and re-executed. The specific steps are as follows: Input the feedback signal into the deep reinforcement learning model, use the feedback signal to perform online learning on the deep reinforcement learning model, and adjust the parameters of the deep reinforcement learning model through the policy gradient method; Based on the updated deep reinforcement learning model, the optimized fire-fighting robot's movement path and water pump start-stop instructions are recalculated. The fire-fighting robot readjusts its movement path according to the optimized fire-fighting robot path, and readjusts the injection parameters according to the optimized water pump start-stop control commands after approaching the fire source.

7. The intelligent monitoring and operation method for tunnel road fire fighting according to claim 6, characterized in that: The fire safety threshold is obtained according to the tunnel cross-sectional area, and the fire status is evaluated by the fire safety threshold, and a multi-spectral scanning is started to detect the fire extinguishing situation. The specific steps are as follows: Based on the geometric structure of the tunnel, the air circulation capacity and heat diffusion capacity are calculated. The fire safety threshold is obtained by combining the air circulation capacity and heat diffusion capacity according to the tunnel parameters, fire dynamics theory and tunnel cross-sectional area. Use distributed optical fiber to monitor temperature distribution, smoke concentration, and flame coverage in the tunnel, and compare the collected data with fire safety thresholds to assess the fire status; Based on the fire status, a multi-spectral scan is initiated to conduct a comprehensive scan of the fire area to detect the fire extinguishing status.

8. An intelligent monitoring and operation system for tunnel road fire fighting, based on the intelligent monitoring and operation method for tunnel road fire fighting according to any one of claims 1 to 7, characterized in that: Including, acquisition module, positioning module, instruction generation module, execution module, optimization module and detection module; The acquisition module is used to collect vibration signals through distributed optical fibers on the tunnel sidewalls, and perform wavelet noise reduction processing to extract vibration features to obtain vibration feature vectors; The positioning module is used to input the vibration feature vector into a pre-trained fire classification model to obtain a fire event, and calculate the fire source position through time domain reflectometry positioning; The instruction generation module is used to build a deep reinforcement learning model based on the fire source location and tunnel state parameters to generate the fire robot path and water pump start and stop instructions; The execution module is used to control the fire-fighting robot to move according to the fire-fighting robot path and water pump start and stop instructions, adjust the water pump injection parameters, and collect fire spread data in real time to generate feedback signals; The optimization module is used to update the reinforcement learning model parameters according to the feedback signal, dynamically optimize the firefighting robot path and the water pump start and stop instructions, and re-execute them; The detection module is used to set a fire safety threshold according to the tunnel cross-sectional area, evaluate the fire status through the fire safety threshold, and detect the fire extinguishing situation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the intelligent monitoring and operation method for tunnel road fire fighting according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent monitoring and operation method for tunnel road fire fighting according to any one of claims 1 to 7 are implemented.

Images