On-site main monitoring device of railway tunnel intelligent fire-fighting monitoring system

Abstract

The invention discloses an on-site main monitoring device for an intelligent fire-fighting monitoring system of a railway tunnel, and relates to the technical field of fire-fighting safety monitoring of the railway tunnel, and the device comprises a data obtaining module which is used for obtaining environment parameters in the tunnel in real time; the edge calculation processing module is used for processing the environmental parameters based on a deep learning algorithm, automatically identifying potential fire risks and predicting a fire behavior development trend; the emergency linkage control module is used for being linked with fire extinguishing equipment in the tunnel when it is judged that the fire risk exists, and automatic emergency disposal is achieved; and the wireless communication module supports various communication protocols, realizes data interaction with a remote monitoring center, and realizes remote monitoring. The system is high in real-time performance, high in intelligent level, high in automation degree and flexible and reliable in communication.

Description

Technical Field

[0001] The present invention relates to the technical field of fire safety monitoring of railway tunnels, and more particularly to an on-site main monitoring device of an intelligent fire monitoring system of a railway tunnel. Background Art

[0002] With the rapid development of railway transportation network, the number and length of railway tunnels are increasing, and their fire safety issues are receiving increasing attention. Traditional railway tunnel fire monitoring systems mainly rely on manual inspections or simple sensor alarm mechanisms, which have problems such as slow response speed, high false alarm rate, and low intelligence. In addition, the on-site monitoring devices of traditional systems have single functions and cannot achieve multi-dimensional data collection and analysis, making it difficult to meet the fire protection needs in the complex environment of modern railway tunnels.

[0003] Therefore, how to provide a new type of on-site main monitoring device for an intelligent fire monitoring system for railway tunnels is a problem that technical personnel in this field need to solve urgently. Summary of the invention

[0004] In view of this, the present invention provides an on-site main monitoring device of an intelligent fire monitoring system for a railway tunnel, which has strong real-time performance, high intelligence level, high degree of automation, and flexible and reliable communication.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] The on-site main monitoring device of the railway tunnel intelligent fire monitoring system includes:

[0007] Data acquisition module: used to obtain environmental parameters in the tunnel in real time;

[0008] Edge computing processing module: used to process environmental parameters based on deep learning algorithms, automatically identify potential fire risks and predict fire development trends;

[0009] Emergency linkage control module: used to link with the fire-fighting equipment in the tunnel when there is a fire risk, to achieve automated emergency response;

[0010] Wireless communication module: supports multiple communication protocols, realizes data interaction with the remote monitoring center, and realizes remote monitoring.

[0011] Preferably, the environmental parameters include temperature, smoke concentration and gas composition.

[0012] Preferably, the edge computing processing module includes:

[0013] Data preprocessing unit: used to normalize environmental parameters and remove noise, and construct spatial distribution maps and time series data;

[0014] CNN model processing unit: used to extract the spatial distribution map based on the CNN model to determine whether there is a fire risk and the risk level;

[0015] Fire development prediction unit: used to process time series data based on the LSTM model and predict the fire development trend.

[0016] Preferably, the CNN model includes an input layer, multiple convolutional layers, an activation function, a pooling layer, a fully connected layer and an output layer in sequence;

[0017] The specific processing process of the CNN model processing unit is:

[0018] The spatial distribution map is converted into a two-dimensional matrix through the input layer;

[0019] Extract local spatial features through multiple convolutional layers;

[0020] After passing through the activation function, pooling layer and fully connected layer in sequence, it is determined whether there is a fire risk and the risk level, and the result is output through the output layer.

[0021] Preferably, the loss function L of the CNN model is total for:

[0022]

[0023] Where n is the number of categories, w i is the weight of the i-th category, y i is the value of the i-th category in the true label, is the model’s predicted probability for this category, Represents all weight parameters w j The square sum of is weighted summation, and λ is the regularization strength control parameter.

[0024] Preferably, the LSTM model includes an input layer, multiple LSTM layers and a fully connected layer, wherein the output of each LSTM layer serves as the input of the next LSTM layer;

[0025] The specific processing process of the fire development prediction unit is as follows:

[0026] Receive time series data through the input layer;

[0027] Multiple LSTM layers process time series data based on cell states and gating mechanisms;

[0028] The output of the LSTM layer is mapped to the fire development trend through the fully connected layer.

[0029] Preferably, the edge computing processing module also includes:

[0030] Fully connected layer joint training unit: used to merge the output of the CNN model and the output of the LSTM model and make a comprehensive judgment through a fully connected layer to output warning information.

[0031] Preferably, it also includes a visual display and interaction module: used to provide touch operation function, and real-time display of environmental parameters, equipment status and warning information in the tunnel.

[0032] Preferably, the wireless communication module includes:

[0033] Network status monitoring unit: used to continuously collect information on network topology and link quality;

[0034] Path evaluation and selection unit: used to evaluate all available paths based on network topology and link quality information and select the best path for data transmission;

[0035] Path switching mechanism unit: When a problem is detected on the existing path or a better new path exists, it quickly switches to the new path to reduce service interruption time.

[0036] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses an on-site main monitoring device for an intelligent fire monitoring system for a railway tunnel, and the present invention has the following advantages:

[0037] 1) Real-time and accuracy

[0038] The data acquisition module can obtain real-time environmental parameters in the tunnel, including but not limited to temperature, smoke concentration, gas composition and flame characteristics. This comprehensive monitoring capability ensures high sensitivity to environmental changes in the tunnel and can capture any abnormal situation at the first time. Compared with traditional manual inspection or simple sensor alarm methods, this method greatly improves the speed and accuracy of detection.

[0039] 2) Intelligent analysis and prediction capabilities

[0040] The edge computing processing module uses technology based on deep learning algorithms to process the collected environmental parameters, which can not only automatically identify potential fire risks, but also further predict the development trend of the fire. This means that the system can not only detect existing problems, but also warn of possible dangers in advance, thus providing a valuable time window for taking preventive measures. In addition, by performing preliminary data processing locally, the need to transmit large amounts of raw data to remote servers is reduced, reducing the pressure on network bandwidth and speeding up response time.

[0041] 3) Automated emergency response function

[0042] The design of the emergency linkage control module enables the system to start immediately and link with the fire-fighting equipment in the tunnel to achieve automated emergency response when it determines that there is a fire risk. This feature greatly shortens the time interval between discovering a problem and taking action, helping to quickly control the fire and reduce losses. At the same time, since the entire process can be completed without human intervention, it can operate effectively even at night or other non-working hours, improving the reliability and efficiency of the system.

[0043] 4) Diversified communication support and high reliability

[0044] The wireless communication module supports a variety of modern communication protocols (such as LoRa, NB-IoT, 5G, etc.), which not only ensures stable data interaction with the remote monitoring center, but also allows the selection of the most suitable communication method according to the specific application scenario. For example, in remote areas or places with poor signal coverage, a long-distance and low-power solution such as LoRa can be selected; and in cases where high-speed and large-volume data transmission is required, the advantages of 5G technology can be used. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0046] Figure 1 This is a block diagram of the on-site main monitoring device of the railway tunnel intelligent fire monitoring system provided by the present invention.

[0047] Figure 2 This is a block diagram of the edge computing processing module provided by the present invention.

[0048] Figure 3 This is a structural diagram of the model provided by the present invention.

[0049] Figure 4 This is a block diagram of the wireless communication module provided by the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] The embodiment of the present invention discloses a railway tunnel intelligent fire monitoring system on-site main monitoring device, such as Figure 1 As shown, including:

[0052] Data acquisition module: used to obtain the environmental parameters in the tunnel in real time; environmental parameters include but are not limited to temperature, smoke concentration, and gas composition.

[0053] Edge computing processing module: used to process environmental parameters based on deep learning algorithms, automatically identify potential fire risks and predict fire development trends;

[0054] Emergency linkage control module: used to link with the fire-fighting equipment in the tunnel when there is a fire risk, to achieve automated emergency response;

[0055] Wireless communication module: supports multiple communication protocols, such as LoRa, NB-IoT, 5G, etc., to realize data interaction with the remote monitoring center and realize remote monitoring.

[0056] In this embodiment, if Figure 2 As shown, the edge computing processing module includes:

[0057] Data preprocessing unit: used to normalize environmental parameters and remove noise, and construct spatial distribution maps and time series data;

[0058] CNN model processing unit: used to extract the spatial distribution map based on the CNN model to determine whether there is a fire risk and the risk level;

[0059] Fire development prediction unit: used to process time series data based on the LSTM model and predict the fire development trend.

[0060] In this embodiment, if Figure 3 As shown, the CNN model includes an input layer, multiple convolutional layers, activation functions, pooling layers, fully connected layers, and output layers in sequence;

[0061] The specific processing process of the CNN model processing unit is:

[0062] The spatial distribution map is converted into a two-dimensional matrix through the input layer; where the rows represent the positions in the tunnel length direction, the columns represent the lateral positions, and the number of channels corresponds to different sensor types.

[0063] Extract local spatial features through multiple convolutional layers, such as local high temperature areas, smoke diffusion directions, etc.

[0064] After passing through the activation function, pooling layer and fully connected layer in sequence, it is determined whether there is a fire risk and the risk level, and the result is output through the output layer.

[0065] The loss function L of the CNN modeltotal for:

[0066]

[0067] Where n is the number of categories, w i is the weight of the i-th category, y i is the value of the i-th category in the true label, is the model’s predicted probability for this category, Represents all weight parameters w j The square sum of λ is weighted and λ controls the strength of regularization.

[0068] The present invention takes into account that different types of errors may have different severities. In the railway tunnel fire monitoring system, the impact of false positives and false negatives may be different. For example, false positives may lead to serious safety accidents, while false positives may cause certain waste of resources, but the relative safety risk is relatively small. Therefore, a weighting mechanism is added to the loss function of the present invention to adjust the weights between different categories. At the same time, in order to avoid overfitting and enhance the generalization ability of the model, a regularization term is added to the loss function.

[0069] The LSTM model includes an input layer, multiple LSTM layers, and a fully connected layer, wherein the output of each LSTM layer is used as the input of the next LSTM layer, and higher-level features can be extracted layer by layer. In this embodiment, three LSTM layers can be set to enhance the ability to capture complex time patterns;

[0070] The specific processing process of the fire development prediction unit is as follows:

[0071] The input layer receives time series data, such as the temperature and smoke concentration in the past 5 minutes. The specific form is:

[0072] Time series data is sampled by time steps (such as one time step every 10 seconds) and forms a sliding window. For example, there are 30 time steps in the past 5 minutes; each time step contains multiple variables (such as temperature, smoke concentration, CO concentration, etc.), so the shape of the input tensor is (number of time steps, number of variables).

[0073] Multiple LSTM layers process time series data based on cell states and gating mechanisms to capture long-term dependencies in time series, such as the transition from slow temperature rise to deflagration in the early stages of a fire. Specifically, cell state: memorizes long-term dependencies (such as the transition from slow temperature rise to deflagration in the early stages of a fire). Gating mechanisms (input gate, forget gate, output gate): dynamically control information retention and forgetting. For example: the input gate captures the rate of temperature rise, and the forget gate ignores short-term fluctuation noise.

[0074] The output of the LSTM layer is mapped to the fire development trend through the fully connected layer. The specific output form is:

[0075] Spread speed (regression task): Use mean squared error loss (MSE) to measure the difference between the predicted value and the true value.

[0076] Spread direction (classification task): The probability distribution of different directions is calculated through the Softmax function to predict the most likely spread direction of the fire (left / right / bidirectional spread).

[0077] Impact range (time series forecast): predict the temperature / smoke distribution at the next time step.

[0078] The loss function of the LSTM model consists of the mean square error (MSE) loss of the regression task and the cross entropy loss of the classification task. The calculation formula of the mean square error (MSE) loss of the regression task is:

[0079]

[0080] Where N is the number of samples, y k is the true value of the kth sample, is the model's predicted value for the kth sample;

[0081] The cross entropy loss calculation formula for classification tasks is:

[0082]

[0083] Among them, C is the number of fire spread direction categories, including left, right, and bidirectional categories, y kc is the true label of the kth sample belonging to the cth category, is the probability that the kth sample belongs to the cth category predicted by the model.

[0084] The edge computing processing module also includes:

[0085] Fully connected layer joint training unit: used to merge the output of the CNN model and the output of the LSTM model and make a comprehensive judgment through a fully connected layer to output warning information.

[0086] The present invention combines the output results of the two models in the fully connected layer to generate more accurate and forward-looking early warning information, working together to significantly improve the performance of the entire system.

[0087] In this embodiment, a visual display and interaction module is also included: it is used to provide touch operation functions and display the environmental parameters, equipment status and warning information in the tunnel in real time.

[0088] In a tunnel environment, the changing trends of key indicators such as temperature, smoke concentration, CO concentration, as well as the working status of the equipment and any possible early warning signals can be displayed through a graphical interface on a large screen or mobile device, making it convenient for on-site staff to quickly view and handle abnormal situations.

[0089] In this embodiment, if Figure 4 As shown, the wireless communication module can automatically optimize the communication path according to the current network conditions to ensure efficient data transmission even under poor network conditions.

[0090] Specifically include:

[0091] Network status monitoring unit: used to continuously collect information on network topology and link quality, such as signal strength, bit error rate, bandwidth utilization, etc.

[0092] Path evaluation and selection unit: Uses a specific algorithm (such as a distance vector algorithm or a link state algorithm) to evaluate all available paths and select the optimal path for data transmission.

[0093] Path switching mechanism unit: When a problem is detected on the existing path or a better new path exists, it quickly switches to the new path to reduce service interruption time.

[0094] The present invention adopts dynamic routing selection technology to automatically optimize the communication path according to the current network status to ensure smooth communication under extreme conditions.

[0095] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0096] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system is characterized by: include: Data acquisition module: used to obtain environmental parameters in the tunnel in real time; Edge computing processing module: used to process environmental parameters based on deep learning algorithms, automatically identify potential fire risks and predict fire development trends; Emergency linkage control module: used to link with the fire-fighting equipment in the tunnel when there is a fire risk, to achieve automated emergency response; Wireless communication module: supports multiple communication protocols, realizes data interaction with the remote monitoring center, and realizes remote monitoring.

2. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 1 is characterized in that: Environmental parameters include temperature, smoke concentration and gas composition.

3. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 1 is characterized in that: The edge computing processing module includes: Data preprocessing unit: used to normalize environmental parameters and remove noise, and construct spatial distribution maps and time series data; CNN model processing unit: used to extract the spatial distribution map based on the CNN model to determine whether there is a fire risk and the risk level; Fire development prediction unit: used to process time series data based on the LSTM model and predict the fire development trend.

4. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 3 is characterized in that: The CNN model consists of an input layer, multiple convolutional layers, activation functions, pooling layers, fully connected layers, and output layers in sequence; The specific processing process of the CNN model processing unit is: The spatial distribution map is converted into a two-dimensional matrix through the input layer; Extract local spatial features through multiple convolutional layers; After passing through the activation function, pooling layer and fully connected layer in sequence, it is determined whether there is a fire risk and the risk level, and the result is output through the output layer.

5. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 4 is characterized in that: The loss function L of the CNN model total for: Where n is the number of categories, w i is the weight of the i-th category, y i is the value of the i-th category in the true label, is the model's predicted probability for this category, Represents all weights w j The square sum of is weighted summation, and λ is the regularization strength control parameter.

6. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 4 is characterized in that: The LSTM model consists of an input layer, multiple LSTM layers, and a fully connected layer, where the output of each LSTM layer serves as the input of the next LSTM layer; The specific processing process of the fire development prediction unit is as follows: Receive time series data through the input layer; Multiple LSTM layers process time series data based on cell states and gating mechanisms; The output of the LSTM layer is mapped to the fire development trend through the fully connected layer.

7. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 6 is characterized in that: The edge computing processing module also includes: Fully connected layer joint training unit: used to merge the output of the CNN model and the output of the LSTM model, make a comprehensive judgment through a fully connected layer, and output warning information.

8. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 7 is characterized in that: It also includes a visual display and interaction module: used to provide touch operation functions, and real-time display of environmental parameters, equipment status and warning information in the tunnel.

9. The on-site main monitoring device of the railway tunnel intelligent fire monitoring system according to claim 1 is characterized in that: The wireless communication module includes: Network status monitoring unit: used to continuously collect information on network topology and link quality; Path evaluation and selection unit: used to evaluate all available paths based on network topology and link quality information and select the best path for data transmission; Path switching mechanism unit: When a problem is detected on the existing path or a better new path exists, it quickly switches to the new path to reduce service interruption time.

Images