Fire prediction method, device, equipment and storage medium for train moving in tunnel

Abstract

The present invention provides a method, device, equipment and storage medium for predicting fires of trains moving in a tunnel, and relates to the field of fire information processing technology, including obtaining information, simulating different fire scenes of trains in train tunnels under different working conditions based on a fire dynamics simulator according to the information, and obtaining simulation data; training a preset fire intelligent identification network model and a temperature distribution advance prediction network model according to the simulation data; inputting the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and outputting the fire source power, fire source location and train speed at the future moment; inputting the temperature data monitored by the temperature sensor at the current moment into the trained temperature distribution advance prediction network model, and outputting the temperature distribution data at the future moment. In the present invention, fire information and future evolution trends of fire can be obtained in real time according to the prediction results, and continuous prediction of temperature distribution can be achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of fire information processing, and in particular to a method, device, equipment and storage medium for predicting fire of a moving train in a tunnel. Background Art

[0002] Most of the existing research on fire adopts the pre-management method, that is, simulating the fire through simulation software and proposing emergency plans based on the simulation results. After the fire occurs, emergency measures are taken according to the emergency plan. Although this type of method can reduce certain casualties and property losses, it faces the disadvantages of poor flexibility, inability to make adjustments according to the on-site situation, and weak on-site command capabilities during fires. Summary of the invention

[0003] The purpose of the present invention is to provide a method, device, equipment and storage medium for predicting fire of moving trains in tunnels to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0004] In a first aspect, the present application provides a method for predicting fire of a moving train in a tunnel, comprising:

[0005] Acquiring information, the information including structural dimensions of the train and the train tunnel, installation positions of temperature sensors on the lateral evacuation platform of the train tunnel, fire source positions, fire source power, and the travel speed of the train in the train tunnel;

[0006] According to the information, different fire scenes of the train in the train tunnel under different working conditions are simulated based on a fire dynamics simulator to obtain simulation data, where the simulation data is time series data;

[0007] According to the simulation data, a preset fire intelligent identification network model and a temperature distribution advance prediction network model are trained, wherein the fire intelligent identification network model includes a long short-term memory network and a fully connected layer, and the temperature distribution advance prediction network model includes a long short-term memory network and a conditional generative adversarial network;

[0008] Input the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and output the fire source power, the fire source location and the train speed at the future moment;

[0009] The temperature data monitored by the temperature sensor at the current moment is input into the trained temperature distribution advance prediction network model, and the temperature distribution data at the future moment is output.

[0010] In a second aspect, the present application also provides a fire prediction device for a moving train in a tunnel, characterized in that it comprises:

[0011] an acquisition module, the acquisition module being used to acquire information, the information including structural dimensions of the train and the train tunnel, installation positions of temperature sensors on the lateral evacuation platform of the train tunnel, fire source positions, fire source power, and the travel speed of the train in the train tunnel;

[0012] A simulation module, wherein the simulation module is used to simulate different fire scenes of the train in the train tunnel under different working conditions based on a fire dynamics simulator according to the information to obtain simulation data, wherein the simulation data is time series data;

[0013] A training module, wherein the training module is used to train a preset fire intelligent identification network model and a temperature distribution advance prediction network model according to the simulation data, wherein the fire intelligent identification network model includes a long short-term memory network and a fully connected layer, and the temperature distribution advance prediction network model includes a long short-term memory network and a conditional generative adversarial network;

[0014] A first prediction module, the first prediction module is used to input the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and output the fire source power, the fire source location and the train speed at a future moment;

[0015] The second prediction module is used to input the temperature data monitored by the temperature sensor at the current moment into the trained temperature distribution advance prediction network model, and output the temperature distribution data at a future moment.

[0016] In a third aspect, the present application also provides a fire prediction device for a moving train in a tunnel, comprising:

[0017] Memory for storing computer programs;

[0018] A processor is used to implement the steps of the method for predicting fires of trains moving in a tunnel when executing the computer program.

[0019] In a fourth aspect, the present application further provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned method for predicting fire based on a moving train in a tunnel are implemented.

[0020] The beneficial effects of the present invention are as follows: the present invention simulates various fire situations of moving trains in tunnels, collects data and trains networks, thereby realizing the recognition of fire information and predicting future fire development trends, so that staff can make correct emergency measures according to fire information and future fire situations, thereby reducing casualties in the tunnel.

[0021] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 It is a schematic flow chart of a method for predicting a fire of a moving train in a tunnel according to an embodiment of the present invention;

[0024] Figure 2 A schematic diagram of a fire scene simulated in an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of the network architecture of a fire intelligent identification network model in an embodiment of the present invention;

[0026] Figure 4 A schematic diagram of the network architecture of a generator in a temperature distribution advance prediction network model according to an embodiment of the present invention;

[0027] Figure 5 Schematic diagram of the network architecture of the discriminator in the temperature distribution advance prediction network model in an embodiment of the present invention;

[0028] Figure 6 It is a schematic diagram of the structure of a fire prediction device for a train moving in a tunnel described in an embodiment of the present invention.

[0029] Markings in the figure: 1. Simulation tunnel; 2. Simulation train; 3. Temperature longitudinal section; 800. Fire prediction equipment for trains moving in tunnel; 801. Processor; 802. Memory; 803. Multimedia component; 804. I / O interface; 805. Communication component. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0031] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0032] Embodiment 1:

[0033] This embodiment provides a method for predicting fire of a moving train in a tunnel.

[0034] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400, and step S500.

[0035] Step S100: Acquiring information, the information including structural dimensions of the train and the train tunnel, installation positions of temperature sensors on the lateral evacuation platform of the train tunnel, fire source positions, fire source power, and the travel speed of the train in the train tunnel;

[0036] Step S200: According to the information, different fire scenes of the train in the train tunnel under different working conditions are simulated based on a fire dynamics simulator to obtain simulation data, where the simulation data is time series data;

[0037] The fire dynamics simulator may be FDS fire simulation software, or OpenFOAM, or other CFD software.

[0038] The step S200 specifically includes:

[0039] According to the structural dimensions of the train and the train tunnel, simulating the train and the train tunnel based on a fire dynamics simulator;

[0040] A temperature longitudinal section is set at the exact center of the lateral evacuation platform of the simulated train tunnel, and the temperature sensor is set at the intersection of the temperature longitudinal section and the top of the train tunnel. Considering that passengers often use the lateral evacuation platforms on both sides of the tunnel to evacuate during a tunnel fire, the temperature longitudinal section is set at the exact center of the lateral evacuation platform. At the same time, a temperature sensor is set at the intersection of the temperature longitudinal section and the top of the tunnel, with an interval of 2m.

[0041] The running speed of the train simulated by the control simulation changes between constant speed, deceleration and stationary.

[0042] The constructed 3D scene of the train tunnel fire is as follows: Figure 2 As shown, a simulated train 2 is running in a simulated tunnel 1, and a temperature longitudinal section 3 is set at the center of the lateral evacuation platform of the simulated train tunnel.

[0043] The control simulation changes the running speed of the train between a constant speed, a deceleration and a stationary speed, including:

[0044] Setting the openings on both sides of the simulated train tunnel to an "OPEN" attribute;

[0045] Constructing a driving speed control function, the driving speed control function includes an environmental wind speed-time relationship function and an environmental wind speed-driving speed relationship function, the environmental wind speed-time piecewise function includes a uniform speed piecewise function, a deceleration piecewise function and a stationary piecewise function;

[0046] The traveling speed of the simulated train is controlled according to the traveling speed control function.

[0047] The driving speed control function includes:

[0048] The ambient wind speed-time relationship function is as follows:

[0049]

[0050] The relationship function between ambient wind speed and driving speed is as follows:

[0051]

[0052] Where V refers to the ambient wind speed, ν0 refers to the initial ambient wind speed, t refers to the train running time, V train refers to the speed of the train, ξ m Refers to the resistance coefficient of the tunnel excluding the annular area, K m It refers to the piston wind action coefficient, and α refers to the tunnel blocking ratio.

[0053] The present invention utilizes the relative wind speed method to equivalently replace the movement of the train in the tunnel. Specifically, the openings on both sides of the tunnel are set as the 'OPEN' attribute, and a function is set to control the size of the ambient wind speed (&RAMPID='WIND_RAMP_SPEED') and its direction (&RAMPID='WIND_RAMP_DIRECTION') in the fire dynamics simulator. When the train is used as the reference system, the ambient wind pouring into the tunnel port is equivalent to the speed of the train when it is running in the tunnel. By controlling the piecewise function of the ambient wind speed changing with time and the function reflecting the relationship between the ambient wind speed and the running speed of the train, various motion states of the train, such as uniform speed, deceleration and stillness, can be realized, thereby utilizing the fire dynamics simulator and the principle of relative motion to realize the entire emergency response process after the train catches fire in the tunnel.

[0054] The simulation of the above fire scenarios has achieved diversified processing of fire scenarios. Mainly by controlling the fire source location, fire source power and train speed, all fire scenarios of train fire in the tunnel are covered as much as possible. The fire source location is mainly reflected in the different positions of the fire source on the top of the carriage. The fire source power refers to the size of the fire source heat release rate, which expresses the ability of the fire source to release heat. When constructing the simulation data set, multiple fire source powers should be set to cover all fire sources on the top of the train in the tunnel as much as possible. Since there is a possibility of fire at any position of the train in the tunnel, the train may also take different driving speeds when driving to the platform ahead. Therefore, the train speed range when constructing the simulation data set should vary from its rated speed to 0.

[0055] After the simulation, the data required for subsequent intelligent fire identification and advance prediction of temperature distribution are extracted, which are mainly divided into three categories: extracting the temperature section diagram, which represents the temperature distribution of the lateral evacuation platform of the tunnel. Each pixel on the section diagram represents a temperature value in the tunnel space. The temperature information on the temperature section diagram is extracted and converted into a temperature distribution matrix; extracting the temperature value collected by the temperature sensor installed on the top of the tunnel; extracting the fire source power, fire source location and train speed under each fire scenario. The above data are exported at a time interval of one second.

[0056] Step S300: training a preset fire intelligent identification network model and a temperature distribution advance prediction network model according to the simulation data, wherein the fire intelligent identification network model includes a long short-term memory network and a fully connected layer, and the temperature distribution advance prediction network model includes a long short-term memory network and a conditional generative adversarial network;

[0057] The step S300 specifically includes:

[0058] Converting the fire source power, the fire source position and the travel speed of the train into vectors;

[0059] The temperature data is set as the input of the preset fire intelligent identification network model, the vector is set as the output of the preset fire intelligent identification network model, and the preset fire intelligent identification network model is trained and tested;

[0060] A first loss function is constructed, and the trained and tested fire intelligent identification network model is iteratively optimized through the first loss function to obtain the trained fire intelligent identification network model.

[0061] Converting the temperature slice diagram into a temperature distribution matrix;

[0062] The temperature data is set as the input of the preset fire intelligent identification network model, the temperature distribution matrix is ​​set as the output of the preset fire intelligent identification network model, and the preset fire intelligent identification network model is trained and tested;

[0063] A second loss function is constructed, and the trained and tested temperature distribution advance prediction network model is iteratively optimized through the second loss function to obtain the trained temperature distribution advance prediction network model.

[0064] A matrix is ​​constructed according to the duration of the temperature sensor collecting the temperature data, the number of the temperature sensors and the temperature data collected by the temperature sensor. For example, the input data of the deep learning model is the temperature data collected by the temperature sensor within 10 seconds. If there are n temperature sensors in total, the input at the i-th moment is A i 10*n Matrix of

[0065] The label values ​​of the deep learning model are divided into two categories, namely, the label values ​​of intelligent fire identification and the label values ​​of advanced prediction of temperature distribution.

[0066] A 1*3 vector is constructed based on the fire source power, the distance of the fire source on the X-axis of the fire dynamics simulator and the speed of the train. For example, the label value of fire intelligent identification is the fire source power, fire source location and train speed corresponding to the working condition. The label value at the i-th moment is the vector B i 1*3 , the vector contains three elements, and the values ​​corresponding to the three elements are fire source power (MW), fire source position (distance on the X axis, m), and train speed (m / s);

[0067] The temperature distribution matrix is ​​constructed according to the number of pixels of the temperature distribution diagram on the X-axis and Z-axis of the fire dynamics simulator. For example, the label value of the temperature distribution at the time i in advance prediction is the temperature distribution matrix C i x*z, x and z represent the number of pixels on the X-axis and Z-axis of the temperature distribution map respectively. The temperature distribution matrix C i x*z The value of each element in represents the temperature value at the corresponding spatial position coordinate.

[0068] Input A at time i i 10*n With label value B i 1*3 , The data sets are in one-to-one correspondence (30 seconds in advance to ensure the advance prediction of temperature distribution) and are randomly divided into training set, verification set and test set in a ratio of 8:1:1. They are subsequently used for the training, verification and testing of the fire intelligent identification network model and the temperature distribution advance prediction network model respectively.

[0069] The fire intelligent identification network model consists of two layers of LSTM network (i.e. long short-term memory network) and FC (i.e. fully connected layer). The network architecture of the fire intelligent identification network model is as follows: Figure 3 The workflow of the fire intelligent identification network model is as follows:

[0070] Temperature sensor data A located at the top of the tunnel within 10 seconds i 10*n (The temperature data within 10 seconds can be the temperature data when the train is in any stage of uniform speed, deceleration and stationary) After passing through the two-layer LSTM network, it is connected to the fully connected neural network. The fully connected network finally outputs a 1*3 vector, which represents the fire source location coordinates, fire source power and train speed respectively.

[0071] The fire source location coordinates, fire source power and train speed output by the network structure are iteratively optimized with the label values ​​of the fire source location, fire source power and train speed through the loss function Loss1 until the loss function becomes stable and no longer decreases, the weights and biases are fixed, and the training ends.

[0072]

[0073] B refers to the fire source power, fire source location and train speed of the label value. pre Refers to the predicted value of fire source power, fire source location and train speed, p is the BB in the training set pre The number of samples.

[0074] The temperature distribution advance prediction network model is mainly composed of LSTM network (i.e. long short-term memory network) and CGAN (i.e. conditional generative adversarial network) network. The workflow of the temperature distribution advance prediction model is as follows:

[0075] like Figure 4As shown in the figure, first, the input (A) in the training set is passed into the generator. The input will first pass through the LSTM network to extract the temperature information in the temperature sensor in time sequence. The adaptive average pooling layer (AdaptiveAvgpool) is immediately after the LSTM to further enrich the information extracted by the LSTM, and convert the output data size of the LSTM into a size suitable for the convolution structure to extract information, and then pass it into the multi-layer convolution (Conv) and deconvolution (DeConv) structure. This part of the structure is the U-net convolution structure. To prevent overfitting, Dropout is set after each convolution layer. Finally, after passing through the adaptive average pooling layer, the information extracted by the U-net structure is converted into a predicted temperature distribution map (C pre );

[0076] like Figure 5 As shown, the input of the discriminator is [A, C] or [A, C pre ], that is, the temperature value within ten seconds (A) obtained by the sensor and the temperature distribution matrix of the evacuation area thirty seconds later (C), or the temperature value within ten seconds (A) and the predicted temperature distribution matrix of the evacuation area thirty seconds later (C pre ), the two are randomly passed into the discriminator. To ensure [A, C] or [A, C pre ] can be spliced ​​together in dimension, using the built-in function reshape in numpy to reshape A, C, C pre After passing through the last convolutional structure of the discriminator, the sigmoid activation function is used to output the discriminator to determine whether the input is [A, C] or [A, C pre ] probability.

[0077] The discriminator determines whether the Nash equilibrium state has been reached. If the return result is NO, the discriminator will act on the generator to further optimize the generator to generate a more realistic temperature distribution; if the return result is YES, the network cannot be further optimized at this time, the training ends, and the model is saved. The loss function of the entire training process is as follows:

[0078]

[0079]

[0080] Among them, A, C, C pre , respectively represent the input (temperature value collected by the temperature sensor within ten seconds), label value (temperature distribution matrix of the passenger evacuation area after thirty seconds), and predicted value (generator output). L1 refers to the mean absolute error loss function (MAE) in Pytorch. It indicates the probability that the discriminator can correctly determine that the input data is a label. When the discriminator correctly determines that the input data is a label, D(C|A) takes the value of 1. represents the probability that the discriminator correctly judges the input data as the predicted value. If the discriminator correctly judges the input data as the predicted value, D(G(C pre |A))) takes the value of 0, and q is CC in the training set pre During the entire training process, the generator is optimized by the discriminator to generate a temperature distribution that is more and more similar to the label. The discriminator also receives the predicted value information from the generator to optimize the discriminator. Through the game and mutual optimization between the generator and the discriminator, the predicted value generated by the generator is almost completely consistent with the label.

[0081] The importance of the output of the fire intelligent identification network model and the temperature distribution advance prediction network model for the fire prediction of trains moving in tunnels is determined based on the weight method.

[0082] Constructing a weighted loss function according to the importance, the first loss function and the second loss function;

[0083] The fire intelligent identification network model and the temperature distribution advance prediction network model are iteratively optimized according to the weighted loss function to obtain a trained temperature distribution advance prediction network model and a trained fire intelligent identification network model.

[0084] The weighted loss function is shown in the formula.

[0085] Loss=ω1Loss1+ω2Loss2

[0086] ω1, ω2 refer to the importance of the fire intelligent identification system and the temperature distribution advance prediction system in the entire system during the model training process, ω1+ω2=1. During the training process, the ratio of ω1 and ω2 can be dynamically adjusted according to the prediction effect of the two systems to ensure that both Loss1 and Loss2 drop to a lower value. If there is no special explanation, ω1=ω2=0.5.

[0087] Step S400: inputting the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and outputting the fire source power, the fire source location and the train speed at the future moment;

[0088] Step S500: input the temperature data monitored by the temperature sensor at the current moment into the trained temperature distribution advance prediction network model, and output the temperature distribution data at the future moment.

[0089] Wherein, the step S400 and the step S500 are performed simultaneously.

[0090] Compared with the prior art, the present invention has the following beneficial effects:

[0091] 1. The system for intelligent identification of fire and advance prediction of temperature distribution of moving trains in tunnels proposed in the present invention firstly intelligently identifies fire information after a train catches fire in a tunnel, so as to ensure that the staff can take correct emergency measures according to the scale of the fire; on the basis of intelligent identification of fire information, the system further predicts the temperature distribution of the lateral evacuation platform during the whole process of the train idling to the platform ahead, and analyzes the safety of the evacuation platform during the train's movement and after parking in advance.

[0092] 2. The method for intelligent fire identification and advance prediction of temperature distribution proposed in the present invention takes very short time, and in practice, fire information and future evolution trends of the fire can be obtained in real time based on the prediction results.

[0093] 3. The temperature distribution advance prediction system in the fire intelligent identification and temperature distribution advance prediction system proposed by the present invention can realize continuous prediction. Since the required input of the temperature distribution advance prediction system is the temperature value collected by the temperature sensor arranged on the lateral evacuation platform, and the output value of the temperature distribution advance prediction system is the temperature distribution of the lateral evacuation platform 30 seconds later, the temperature value of the position corresponding to the input in the predicted temperature distribution map can be taken and then transmitted to the temperature distribution advance prediction system again, and the temperature distribution of the lateral evacuation platform 30 seconds later can be output again, so as to realize continuous prediction of temperature distribution.

[0094] Embodiment 2:

[0095] This embodiment provides a device for predicting fire of a moving train in a tunnel, the device comprising:

[0096] an acquisition module, the acquisition module being used to acquire information, the information including structural dimensions of the train and the train tunnel, installation positions of temperature sensors on the lateral evacuation platform of the train tunnel, fire source positions, fire source power, and the travel speed of the train in the train tunnel;

[0097] A simulation module, wherein the simulation module is used to simulate different fire scenes of the train in the train tunnel under different working conditions based on a fire dynamics simulator according to the information to obtain simulation data, wherein the simulation data is time series data;

[0098] A training module, wherein the training module is used to train a preset fire intelligent identification network model and a temperature distribution advance prediction network model according to the simulation data, wherein the fire intelligent identification network model includes a long short-term memory network and a fully connected layer, and the temperature distribution advance prediction network model includes a long short-term memory network and a conditional generative adversarial network;

[0099] A first prediction module, the first prediction module is used to input the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and output the fire source power, the fire source location and the train speed at a future moment;

[0100] The second prediction module is used to input the temperature data monitored by the temperature sensor at the current moment into the trained temperature distribution advance prediction network model, and output the temperature distribution data at a future moment.

[0101] The simulation module comprises:

[0102] A simulation module, wherein the simulation module is used to simulate the train and the train tunnel based on a fire dynamics simulator according to the structural dimensions of the train and the train tunnel;

[0103] A setting module, the setting module is used to set a temperature longitudinal section at the center of the lateral evacuation platform of the simulated train tunnel, and set the temperature sensor at the intersection of the temperature longitudinal section and the top of the train tunnel;

[0104] A control module is used to control the running speed of the simulated train to change between a constant speed, a deceleration and a stationary state.

[0105] The simulation data includes the temperature data monitored by the temperature sensor, the fire source power, the fire source location and the travel speed of the train, and the training module includes:

[0106] A first conversion module, the first conversion module is used to convert the fire source power, the fire source position and the travel speed of the train into a vector;

[0107] A first training and testing module, the training and testing module is used to set the temperature data as the input of the preset fire intelligent identification network model, set the vector as the output of the preset fire intelligent identification network model, and train and test the preset fire intelligent identification network model;

[0108] The first building module is used to construct a first loss function, and iteratively optimize the trained and tested fire intelligent identification network model through the first loss function to obtain the trained fire intelligent identification network model.

[0109] The simulation data also includes a temperature section diagram of the lateral evacuation platform of the train tunnel, and the training module includes:

[0110] A second conversion module, the second conversion module is used to convert the temperature section diagram into a temperature distribution matrix;

[0111] A second training and testing module, the second training and testing module is used to set the temperature data as the input of the preset fire intelligent identification network model, set the temperature distribution matrix as the output of the preset fire intelligent identification network model, and train and test the preset fire intelligent identification network model;

[0112] The second construction module is used to construct a second loss function, and iteratively optimize the temperature distribution advance prediction network model after training and testing through the second loss function to obtain the trained temperature distribution advance prediction network model.

[0113] It should be noted that, regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0114] Embodiment 3:

[0115] Corresponding to the above method embodiment, this embodiment also provides a fire prediction device for a train moving in a tunnel. The fire prediction device for a train moving in a tunnel described below and the fire prediction method for a train moving in a tunnel described above can refer to each other.

[0116] Figure 6 FIG. 8 is a block diagram of a fire prediction device 800 for a train moving in a tunnel according to an exemplary embodiment. Figure 6 As shown, the fire prediction device 800 for trains moving in a tunnel may include: a processor 801 and a memory 802. The fire prediction device 800 for trains moving in a tunnel may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0117] The processor 801 is used to control the overall operation of the in-tunnel moving train fire prediction device 800 to complete all or part of the steps in the in-tunnel moving train fire prediction method. The memory 802 is used to store various types of data to support the operation of the in-tunnel moving train fire prediction device 800, and these data may include, for example, instructions for any application or method operating on the in-tunnel moving train fire prediction device 800, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 802 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 memory, flash memory, disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be keyboards, mice, buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the fire prediction device 800 for moving trains in the tunnel and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 can include: Wi-Fi module, Bluetooth module, NFC module.

[0118] In an exemplary embodiment, the fire prediction device 800 for a train moving in a tunnel can be implemented by one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned method for predicting fires for trains moving in a tunnel.

[0119] In another exemplary embodiment, a computer storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned method for predicting fires of trains moving in a tunnel are implemented. For example, the computer storage medium may be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions may be executed by the processor 801 of the device 800 for predicting fires of trains moving in a tunnel to complete the above-mentioned method for predicting fires of trains moving in a tunnel.

[0120] Embodiment 4:

[0121] Corresponding to the above method embodiment, a storage medium is also provided in this embodiment. The storage medium described below and the method for predicting a fire of a moving train in a tunnel described above can refer to each other.

[0122] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for predicting a fire of a moving train in a tunnel of the above method embodiment.

[0123] The storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other storage medium that can store program codes.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0125] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for predicting fire of a moving train in a tunnel, characterized in that: include: Acquiring information, the information including structural dimensions of the train and the train tunnel, installation positions of temperature sensors on the lateral evacuation platform of the train tunnel, fire source positions, fire source power, and the travel speed of the train in the train tunnel; setting a temperature longitudinal section at the exact center of the lateral evacuation platform of the simulated train tunnel, and setting the temperature sensor at the intersection of the temperature longitudinal section and the top of the train tunnel; According to the information, different fire scenes of the train in the train tunnel under different working conditions are simulated based on a fire dynamics simulator to obtain simulation data, where the simulation data is time series data; According to the simulation data, a preset fire intelligent identification network model and a temperature distribution advance prediction network model are trained, wherein the fire intelligent identification network model includes a long short-term memory network and a fully connected layer, and the temperature distribution advance prediction network model includes a long short-term memory network and a conditional generative adversarial network; Input the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and output the fire source power, the fire source location and the train speed at the future moment; The temperature data monitored by the temperature sensor at the current moment is input into the trained temperature distribution advance prediction network model, and the temperature distribution data at the future moment is output.

2. The method for predicting fire of a moving train in a tunnel according to claim 1, characterized in that ,According to the information, different fire scenarios of the train in the train tunnel under different working conditions are simulated based on a fire dynamics simulator, including: According to the structural dimensions of the train and the train tunnel, simulating the train and the train tunnel based on a fire dynamics simulator; The running speed of the train simulated by the control simulation changes between constant speed, deceleration and stationary.

3. The method for predicting fire of a moving train in a tunnel according to claim 2 is characterized in that The simulation data includes the temperature data monitored by the temperature sensor, the fire source power, the fire source location and the travel speed of the train. The preset fire intelligent identification network model and the temperature distribution advance prediction network model are trained according to the simulation data, including: Converting the fire source power, the fire source position and the travel speed of the train into vectors; The temperature data is set as the input of the preset fire intelligent identification network model, the vector is set as the output of the preset fire intelligent identification network model, and the preset fire intelligent identification network model is trained and tested; A first loss function is constructed, and the trained and tested fire intelligent identification network model is iteratively optimized through the first loss function to obtain the trained fire intelligent identification network model.

4. The method for predicting fire of a moving train in a tunnel according to claim 3 is characterized in that The simulation data also includes a temperature cross-sectional diagram of the lateral evacuation platform of the train tunnel. The preset fire intelligent identification network model and temperature distribution advance prediction network model are trained according to the simulation data, including: Converting the temperature slice diagram into a temperature distribution matrix; The temperature data is set as the input of the preset fire intelligent identification network model, the temperature distribution matrix is ​​set as the output of the preset fire intelligent identification network model, and the preset fire intelligent identification network model is trained and tested; A second loss function is constructed, and the trained and tested temperature distribution advance prediction network model is iteratively optimized through the second loss function to obtain the trained temperature distribution advance prediction network model.

5. A fire prediction device for trains moving in a tunnel, characterized in that: include: An acquisition module, the acquisition module is used to acquire information, the information including the structural dimensions of the train and the train tunnel, the installation position of the temperature sensor on the lateral evacuation platform of the train tunnel, the fire source position, the fire source power and the running speed of the train in the train tunnel; a temperature longitudinal section is set at the center of the lateral evacuation platform of the simulated train tunnel, and the temperature sensor is set at the intersection of the temperature longitudinal section and the top of the train tunnel; A simulation module, wherein the simulation module is used to simulate different fire scenes of the train in the train tunnel under different working conditions based on a fire dynamics simulator according to the information to obtain simulation data, wherein the simulation data is time series data; A training module, wherein the training module is used to train a preset fire intelligent identification network model and a temperature distribution advance prediction network model according to the simulation data, wherein the fire intelligent identification network model includes a long short-term memory network and a fully connected layer, and the temperature distribution advance prediction network model includes a long short-term memory network and a conditional generative adversarial network; A first prediction module, the first prediction module is used to input the temperature data monitored by the temperature sensor at the current moment into the trained fire intelligent identification network model, and output the fire source power, the fire source location and the train speed at a future moment; The second prediction module is used to input the temperature data monitored by the temperature sensor at the current moment into the trained temperature distribution advance prediction network model, and output the temperature distribution data at a future moment.

6. The fire prediction device for trains moving in a tunnel according to claim 5, characterized in that: The simulation module comprises: A simulation module, wherein the simulation module is used to simulate the train and the train tunnel based on a fire dynamics simulator according to the structural dimensions of the train and the train tunnel; A setting module, the setting module is used to set a temperature longitudinal section at the center of the lateral evacuation platform of the simulated train tunnel, and set the temperature sensor at the intersection of the temperature longitudinal section and the top of the train tunnel; A control module is used to control the running speed of the simulated train to change between a constant speed, a decelerated speed and a stationary speed.

7. The fire prediction device for trains moving in a tunnel according to claim 5, characterized in that: The simulation data includes the temperature data monitored by the temperature sensor, the fire source power, the fire source location and the travel speed of the train, and the training module includes: A first conversion module, the first conversion module is used to convert the fire source power, the fire source position and the travel speed of the train into a vector; A first training and testing module, the training and testing module is used to set the temperature data as the input of the preset fire intelligent identification network model, set the vector as the output of the preset fire intelligent identification network model, and train and test the preset fire intelligent identification network model; The first building module is used to construct a first loss function, and iteratively optimize the trained and tested fire intelligent identification network model through the first loss function to obtain the trained fire intelligent identification network model.

8. The fire prediction device for trains moving in a tunnel according to claim 7, characterized in that: The simulation data also includes a temperature section diagram of the lateral evacuation platform of the train tunnel, and the training module includes: A second conversion module, the second conversion module is used to convert the temperature section diagram into a temperature distribution matrix; A second training and testing module, the second training and testing module is used to set the temperature data as the input of the preset fire intelligent identification network model, set the temperature distribution matrix as the output of the preset fire intelligent identification network model, and train and test the preset fire intelligent identification network model; The second construction module is used to construct a second loss function, and iteratively optimize the temperature distribution advance prediction network model after training and testing through the second loss function to obtain the trained temperature distribution advance prediction network model.

9. A fire prediction device for trains moving in a tunnel, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the method for predicting a fire of a moving train in a tunnel as described in any one of claims 1 to 4 when executing the computer program.

10. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for predicting a fire of a moving train in a tunnel as claimed in any one of claims 1 to 4.

Images