Vehicle fire warning method and system

Abstract

The present application relates to the field of vehicle fire self-detection technology, specifically disclosing a vehicle fire warning method and system. The method employs deep learning-based artificial intelligence detection technology to capture the dynamic time-series characteristic information of the temperature and pressure of the gas discharged from the gas exhaust portion of the battery pack through a convolutional neural network model, and mines the high-dimensional time-series implicit correlation characteristic distribution information between the two as the input of a decoder. In this way, an accurate warning is provided for a fire in the battery pack housing, and the battery fire risk level is determined, thereby reserving sufficient time to formulate a corresponding self-rescue plan, avoiding serious accidents caused by fire, and protecting the lives of passengers.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle fire self-detection, and more specifically, to a vehicle fire warning method and system. Background Art

[0002] Nickel-cadmium batteries, nickel-metal hydride batteries, nickel-zinc batteries and lithium secondary batteries are currently commercialized secondary batteries. Among them, lithium secondary batteries have attracted much attention because, compared with nickel-based secondary batteries, lithium secondary batteries have almost no memory effect, thereby enabling free discharge, extremely low self-discharge rate, and high energy density. At the same time, when lithium secondary batteries are charged and discharged, lithium secondary batteries generate heat. In the case where heat is not effectively removed but accumulated, battery degradation may occur, and the safety of the battery may also be greatly reduced. In particular, in batteries that require high-speed charging and discharging characteristics, such as power sources for electric vehicles, hybrid electric vehicles, etc., a large amount of heat will be generated and the battery volume will expand in the process of instantaneously providing high output, and in severe cases, fire may occur.

[0003] However, in the event of a fire in the battery pack, the temperature or voltage sensors may be damaged, making it impossible to accurately sense the battery status. Furthermore, in the event of a fire near the battery management system, the battery management system may not function properly, making it impossible to properly implement passenger protection measures.

[0004] To address the above issues, Chinese patent CN113748042A discloses a vehicle battery fire sensing device and method. This device uses a sensor to measure the temperature and pressure of gas discharged from the battery pack's gas exhaust port to determine whether the battery pack housing has caught fire. If a fire occurs in the battery pack housing, the device quickly and accurately notifies passengers of the situation. However, in actual application, it has been found that if a warning is issued after a battery pack housing fire is detected, the safety hazard has already occurred, and even with a warning, there may still be insufficient time to respond.

[0005] Therefore, an optimized vehicle fire warning solution is desired. Summary of the Invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a vehicle fire warning method and system, which adopts artificial intelligence detection technology based on deep learning to capture the dynamic change characteristic information of the temperature and pressure of the gas discharged from the gas discharge part of the battery pack in time series through a convolutional neural network model, and mines the high-dimensional time series implicit correlation feature distribution information between the two as the input of the decoder. In this way, an accurate warning is given for the fire of the battery pack shell, and the battery fire risk level is judged at the same time, so as to reserve sufficient time to make corresponding self-rescue plans, avoid serious accidents caused by fire, and protect the lives of passengers.

[0007] According to one aspect of the present application, a vehicle fire warning method is provided, which includes:

[0008] Obtaining temperature and pressure values ​​of gas generated by the monitored battery pack at a plurality of predetermined time points within a predetermined time period;

[0009] Arranging the temperature values ​​and pressure values ​​at the plurality of predetermined time points into a temperature input vector and a pressure input vector respectively according to the time dimension;

[0010] performing correlation coding on the temperature input vector and the pressure input vector to obtain a temperature-pressure correlation input matrix;

[0011] Passing the temperature-pressure correlation input matrix through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix;

[0012] performing eigenvalue differentiation on the temperature-pressure correlation characteristic matrix to obtain an optimized temperature-pressure correlation characteristic matrix;

[0013] Passing the optimized temperature-pressure correlation feature matrix through a decoder to obtain a decoded value representing a battery fire risk level; and

[0014] Based on the comparison between the decoded value and a predetermined threshold, an early warning prompt is generated.

[0015] In the vehicle fire warning method, the temperature input vector and the pressure input vector are associated and encoded to obtain a temperature-pressure associated input matrix, including: the temperature input vector and the pressure input vector are associated and encoded to obtain a temperature-pressure associated input matrix according to the following formula; wherein the formula is:

[0016]

[0017] in represents the transposed vector of the temperature input vector, V brepresents the pressure input vector, M represents the temperature-pressure correlation input matrix, Represents matrix multiplication.

[0018] In the above-mentioned vehicle fire warning method, the temperature-pressure correlation input matrix is ​​passed through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix, including: using the convolution layer of the first layer of the convolutional neural network model to convolve the input data with a first convolution kernel to generate a first convolution feature map; using the pooling layer of the first layer of the convolutional neural network model to perform pooling processing on the first convolution feature map along the channel dimension to generate a first pooling feature matrix; using the activation layer of the first layer of the convolutional neural network model to activate the first pooling feature matrix to generate a first activation feature map; using the convolutional neural network model to perform activation processing on the first pooling feature matrix to generate a first activation feature map. The convolution layer of the second layer of the type convolutions the input data with a second convolution kernel to generate a second convolution feature map, and the second convolution kernel and the first convolution kernel are transposed of each other; the pooling layer of the second layer of the convolutional neural network model is used to perform pooling processing on the second convolution feature map along the channel dimension to generate a second pooling feature matrix; and the activation layer of the second layer of the convolutional neural network model is used to activate the second pooling feature matrix to generate a second activation feature map; wherein, the output of the last layer of the convolutional neural network model is the temperature-pressure correlation feature matrix, and the input of the first layer of the convolutional neural network model is the temperature-pressure correlation input matrix.

[0019] In the vehicle fire warning method, performing eigenvalue differentiation on the temperature-pressure correlation feature matrix to obtain an optimized temperature-pressure correlation feature matrix includes performing interactive enhancement based on distinguishable physical excitation on the temperature-pressure correlation feature matrix using the following formula to obtain the optimized temperature-pressure correlation feature matrix; wherein the formula is:

[0020]

[0021] Where M is the temperature-pressure correlation feature matrix, a and b are predetermined hyperparameters, and represents positional addition and subtraction of the feature matrix, division represents dividing each position of the feature matrix by the corresponding value, and Cov(·) represents the convolution operation through a single convolutional layer, and X is the optimized temperature-pressure correlation feature matrix.

[0022] In the vehicle fire warning method, the optimized temperature-pressure correlation feature matrix is ​​passed through a decoder to obtain a decoded value representing the battery fire risk level, including: using the decoder to perform decoding regression on the optimized temperature-pressure correlation feature matrix using the following formula to obtain the decoded value; wherein the formula is: Where X is the optimized temperature-pressure correlation feature matrix, Y is the decoded value, and W is the weight matrix. Represents matrix multiplication.

[0023] According to another aspect of the present application, a vehicle fire warning system is provided, comprising:

[0024] A monitoring module, configured to obtain temperature and pressure values ​​of gas generated by the monitored battery pack at a plurality of predetermined time points within a predetermined time period;

[0025] a data structuring module, configured to arrange the temperature values ​​and pressure values ​​at the plurality of predetermined time points into a temperature input vector and a pressure input vector respectively according to a time dimension;

[0026] an association coding module, configured to perform association coding on the temperature input vector and the pressure input vector to obtain a temperature-pressure association input matrix;

[0027] a feature extraction module, configured to pass the temperature-pressure correlation input matrix through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix;

[0028] an eigenvalue distinguishing module, configured to perform eigenvalue distinguishing on the temperature-pressure correlation characteristic matrix to obtain an optimized temperature-pressure correlation characteristic matrix;

[0029] a decoding regression module, configured to pass the optimized temperature-pressure correlation feature matrix through a decoder to obtain a decoded value representing a battery fire risk level; and

[0030] The warning generation module is used to generate a warning prompt based on the comparison between the decoded value and a predetermined threshold.

[0031] In the above vehicle fire warning system, the correlation coding module is further used to: perform correlation coding on the temperature input vector and the pressure input vector using the following formula to obtain a temperature-pressure correlation input matrix; wherein the formula is:

[0032]

[0033] in represents the transposed vector of the temperature input vector, V b represents the pressure input vector, M represents the temperature-pressure correlation input matrix, Represents matrix multiplication.

[0034] In the above-mentioned vehicle fire warning system, the feature extraction module is further used to: use the convolution layer of the first layer of the convolutional neural network model to perform convolution processing on the input data with a first convolution kernel to generate a first convolution feature map; use the pooling layer of the first layer of the convolutional neural network model to perform pooling processing on the first convolution feature map along the channel dimension to generate a first pooling feature matrix; use the activation layer of the first layer of the convolutional neural network model to perform activation processing on the first pooling feature matrix to generate a first activation feature map; use the convolution layer of the second layer of the convolutional neural network model to perform convolution processing on the input data with a second convolution kernel Perform convolution processing to generate a second convolution feature map, where the second convolution kernel and the first convolution kernel are transposed of each other; use the pooling layer of the second layer of the convolutional neural network model to perform pooling processing on the second convolution feature map along the channel dimension to generate a second pooling feature matrix; and use the activation layer of the second layer of the convolutional neural network model to perform activation processing on the second pooling feature matrix to generate a second activation feature map; wherein, the output of the last layer of the convolutional neural network model is the temperature-pressure correlation feature matrix, and the input of the first layer of the convolutional neural network model is the temperature-pressure correlation input matrix.

[0035] In the vehicle fire warning system, the eigenvalue distinguishing module is further configured to perform interactive reinforcement based on distinguishable physical excitation on the temperature-pressure correlation characteristic matrix using the following formula to obtain the optimized temperature-pressure correlation characteristic matrix; wherein the formula is:

[0036]

[0037] Where M is the temperature-pressure correlation feature matrix, a and b are predetermined hyperparameters, and represents positional addition and subtraction of the feature matrix, division represents dividing each position of the feature matrix by the corresponding value, and Cov(·) represents the convolution operation through a single convolutional layer, and X is the optimized temperature-pressure correlation feature matrix.

[0038] In the above vehicle fire warning system, the decoding regression module is further used to: use the decoder to perform decoding regression on the optimized temperature-pressure correlation feature matrix using the following formula to obtain the decoding value; wherein the formula is: Where X is the optimized temperature-pressure correlation feature matrix, Y is the decoded value, and W is the weight matrix. Represents matrix multiplication.

[0039] According to another aspect of the present application, an electronic device is provided, comprising: a processor; and a memory, wherein computer program instructions are stored in the memory, and when the computer program instructions are executed by the processor, the processor executes the vehicle fire warning method as described above.

[0040] According to another aspect of the present application, a computer-readable medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the vehicle fire warning method as described above.

[0041] Compared to existing technologies, the vehicle fire warning method and system provided in this application utilizes deep learning-based artificial intelligence detection technology to capture the dynamic time-series characteristics of the temperature and pressure of the gas discharged from the battery pack's gas exhaust section through a convolutional neural network model. The system also mines the high-dimensional time-series implicit correlation feature distribution information between the two as input to the decoder. This provides an accurate warning of battery pack shell fires and simultaneously determines the battery fire risk level, allowing sufficient time to develop a self-rescue plan, preventing serious accidents caused by fires and protecting the lives of passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0043] Figure 1 This is an application scenario diagram of the vehicle fire warning method according to an embodiment of the present application.

[0044] Figure 2 Flowchart of a vehicle fire warning method according to an embodiment of the present application.

[0045] Figure 3 2 is a schematic diagram of a vehicle fire warning method according to an embodiment of the present application.

[0046] Figure 4 4 is a block diagram of a vehicle fire warning system according to an embodiment of the present application.

[0047] Figure 5 is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0048] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0049] Figure 1 FIG is an application scenario diagram of the vehicle fire warning method according to an embodiment of the present application. Figure 1 As shown, in this application scenario, first, a temperature sensor (e.g. Figure 1 Se1) and a pressure sensor (e.g., as shown in Figure 1 Se2) to obtain the number of predetermined time points within a predetermined time period by the vehicle (for example, Figure 1 The monitored battery pack (e.g., Figure 1 The gas generated by P) (e.g., Figure 1 Furthermore, the temperature and pressure values ​​of the gas generated by the monitored battery pack at multiple predetermined time points within the predetermined time period are input to a server (e.g., Figure 1 In S) as shown, the server is capable of processing the temperature and pressure values ​​of the gas generated by the monitored battery pack at multiple predetermined time points within the predetermined time period based on the vehicle fire warning algorithm to obtain a decoded value for indicating the battery fire risk level, and generating a warning prompt based on a comparison between the decoded value and a predetermined threshold.

[0050] After introducing the basic principles of the present application, various non-limiting embodiments of the present application will be described in detail with reference to the accompanying drawings.

[0051] Exemplary Methods

[0052] Figure 2 FIG. 1 is a flow chart of a vehicle fire warning method according to an embodiment of the present application. Figure 2As shown, the vehicle fire warning method according to an embodiment of the present application includes: S110, obtaining the temperature value and pressure value of the gas generated by the monitored battery pack at multiple predetermined time points within a predetermined time period; S120, arranging the temperature values ​​and pressure values ​​of the multiple predetermined time points into a temperature input vector and a pressure input vector according to the time dimension respectively; S130, performing association encoding on the temperature input vector and the pressure input vector to obtain a temperature-pressure association input matrix; S140, passing the temperature-pressure association input matrix through a convolutional neural network model with mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure association feature matrix; S150, performing eigenvalue differentiation on the temperature-pressure association feature matrix to obtain an optimized temperature-pressure association feature matrix; S160, passing the optimized temperature-pressure association feature matrix through a decoder to obtain a decoded value for indicating the battery fire risk level; and, S170, generating a warning prompt based on a comparison between the decoded value and a predetermined threshold.

[0053] Figure 3 FIG is a structural diagram of a vehicle fire warning method according to an embodiment of the present application. Figure 3 As shown, in this architecture, first, the temperature value and pressure value of the gas generated by the monitored battery pack at multiple predetermined time points within a predetermined time period are obtained; then, the temperature values ​​and pressure values ​​of the multiple predetermined time points are arranged into a temperature input vector and a pressure input vector according to the time dimension respectively; then, the temperature input vector and the pressure input vector are associated and encoded to obtain a temperature-pressure associated input matrix; then, the temperature-pressure associated input matrix is ​​passed through a convolutional neural network model with mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure associated feature matrix; then, the temperature-pressure associated feature matrix is ​​distinguished by eigenvalues ​​to obtain an optimized temperature-pressure associated feature matrix; then, the optimized temperature-pressure associated feature matrix is ​​passed through a decoder to obtain a decoded value for indicating the battery fire risk level; finally, based on the comparison between the decoded value and a predetermined threshold, an early warning prompt is generated.

[0054] As mentioned above, Chinese patent CN113748042A discloses a vehicle battery fire sensing device and method that uses sensors to measure the temperature and pressure of gas discharged from the battery pack's gas exhaust section to determine whether the battery pack housing has caught fire. When a fire occurs in the battery pack housing, the device quickly and accurately notifies passengers of the situation. However, in actual applications, it has been found that if a warning is issued after a fire is detected in the battery pack housing, the safety hazard has already occurred. Even with a warning, there may still be insufficient time to respond. Therefore, an optimized vehicle fire warning solution is desired.

[0055] Accordingly, considering that when actually carrying out a fire warning of the battery pack shell to achieve self-rescue from a fire, in order to reserve sufficient time for response and to avoid serious accidents caused by the fire, it is necessary to effectively extract and capture the dynamic time-series characteristic information of the temperature and pressure of the gas discharged from the gas discharge part of the battery pack, so as to provide early warning of possible safety problems. However, in the actual information extraction process, considering that there is a dynamic time-series correlation change between the temperature and pressure of the gas, the difficulty lies in how to establish a mapping relationship between the time-series change of the temperature value and the time-series change of the pressure value of the gas generated by the battery pack, so as to accurately warn of the fire of the battery pack shell and judge the battery fire risk level at the same time, so as to reserve sufficient time to make corresponding self-rescue plans, avoid serious accidents caused by the fire, and protect the lives of passengers.

[0056] In recent years, deep learning and neural networks have been widely used in fields such as computer vision, natural language processing, and text signal processing. Furthermore, deep learning and neural networks have demonstrated capabilities approaching or even surpassing those of humans in areas such as image classification, object detection, semantic segmentation, and text translation.

[0057] The development of deep learning and neural networks has provided new solutions and solutions for exploring the complex mapping relationship between the time series changes in the temperature values ​​and the time series changes in the pressure values ​​of the gas generated by the battery pack. Those skilled in the art will appreciate that a deep neural network model based on deep learning can be trained using appropriate strategies, such as a gradient descent backpropagation algorithm, to adjust the parameters of the deep neural network model so that it can simulate complex nonlinear relationships between objects. This is clearly suitable for simulating and establishing the complex mapping relationship between the time series changes in the temperature values ​​and the time series changes in the pressure values ​​of the gas generated by the battery pack.

[0058] In step S110, the temperature and pressure values ​​of the gas generated by the monitored battery pack are obtained at multiple predetermined time points within a predetermined time period. The temperature and pressure values ​​of the gas can be obtained by a temperature sensor and a pressure sensor, respectively. Specifically, the temperature value is obtained by a temperature sensor deployed on the monitored battery pack of the vehicle, and / or the pressure value is obtained by a pressure sensor deployed on the monitored battery pack of the vehicle.

[0059] In step S120, the temperature values ​​and pressure values ​​at the plurality of predetermined time points are arranged according to the time dimension into a temperature input vector and a pressure input vector, respectively. Considering that the temperature values ​​and pressure values ​​of the gas generated by the monitored battery pack have dynamic change patterns in the time dimension, and that these two parameter data also have characteristic distribution information of correlation in the time dimension, in order to accurately and timely provide a fire warning for the battery pack shell, in the technical solution of the present application, the temperature values ​​and pressure values ​​at the plurality of predetermined time points are arranged according to the time dimension into a temperature input vector and a pressure input vector, respectively, so as to integrate the parameter distribution information of the temperature values ​​and pressure values ​​of the gas in the time series.

[0060] In step S130, the temperature input vector and the pressure input vector are association-encoded to obtain a temperature-pressure association input matrix. To facilitate the subsequent sufficient and accurate extraction of temporal correlation features between the temperature and pressure values ​​of the gas, the temperature input vector and the pressure input vector are further association-encoded to establish a correlation relationship between the parameter temporal distribution information of the gas temperature and pressure values, thereby obtaining the temperature-pressure association input matrix.

[0061] Specifically, in an embodiment of the present application, the encoding process of performing association encoding on the temperature input vector and the pressure input vector to obtain a temperature-pressure association input matrix includes: performing association encoding on the temperature input vector and the pressure input vector according to the following formula to obtain a temperature-pressure association input matrix; wherein the formula is:

[0062]

[0063] in represents the transposed vector of the temperature input vector, V b represents the pressure input vector, M represents the temperature-pressure correlation input matrix, Represents matrix multiplication.

[0064] In step S140, the temperature-pressure correlation input matrix is ​​passed through a convolutional neural network model with mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix. That is, a convolutional neural network model with excellent performance in extracting local implicit correlation features is used to extract the collaborative correlation features of the temperature and pressure values ​​of the gas. However, considering that the temperature and pressure values ​​of the gas have a considerable degree of correlation in the time dimension, in order to fully extract the temporal collaborative correlation features of the temperature and pressure values ​​of the gas to accurately judge the fire risk level of the battery pack shell, in the technical solution of the present application, the temperature-pressure correlation input matrix is ​​further subjected to feature mining by using a convolutional neural network model with mutually transposed convolution kernels in adjacent layers to extract deeper and more sufficient temporal collaborative correlation implicit feature information of the temperature and pressure of the gas, thereby obtaining a temperature-pressure correlation feature matrix. In particular, here, the adjacent convolutional layers of the convolutional neural network model use mutually transposed convolution kernels to simultaneously update the network parameters and search for the network parameter structure suitable for a specific data structure during training, thereby improving the accuracy of subsequent decoding.

[0065] Specifically, in an embodiment of the present application, the temperature-pressure correlation input matrix is ​​passed through a convolutional neural network model with mutually transposed convolution kernels in adjacent layers to obtain an encoding process of a temperature-pressure correlation feature matrix, including: using the convolution layer of the first layer of the convolutional neural network model to convolve the input data with a first convolution kernel to generate a first convolution feature map; using the pooling layer of the first layer of the convolutional neural network model to perform pooling processing on the first convolution feature map along the channel dimension to generate a first pooling feature matrix; using the activation layer of the first layer of the convolutional neural network model to activate the first pooling feature matrix to generate a first activation feature map; using the convolutional neural network model to perform activation processing on the first pooling feature matrix to generate a first activation feature map. The convolution layer of the second layer of the network model convolves the input data with a second convolution kernel to generate a second convolution feature map, and the second convolution kernel and the first convolution kernel are transposed of each other; the pooling layer of the second layer of the convolutional neural network model is used to perform pooling processing on the second convolution feature map along the channel dimension to generate a second pooling feature matrix; and, the activation layer of the second layer of the convolutional neural network model is used to activate the second pooling feature matrix to generate a second activation feature map; wherein, the output of the last layer of the convolutional neural network model is the temperature-pressure correlation feature matrix, and the input of the first layer of the convolutional neural network model is the temperature-pressure correlation input matrix.

[0066] In step S150, the temperature-pressure correlation feature matrix is ​​distinguished by eigenvalues ​​to obtain an optimized temperature-pressure correlation feature matrix. In particular, in the technical solution of the present application, a specific data structure in the temperature-pressure correlation input matrix is ​​captured by using adjacent layers using a combination structure of mutually transposed convolution kernels as a specific data block searcher to obtain the temperature-pressure correlation feature matrix. In other words, the importance of the eigenvalues ​​of each position in the specific data structure in the temperature-pressure correlation feature matrix exceeds that of the eigenvalues ​​at other positions. Therefore, if the eigenvalues ​​in the temperature-pressure correlation feature matrix can be effectively distinguished based on the eigenvalue positions in the decoding regression task, it is obvious that the training speed of the model and the accuracy of the decoding value can be improved.

[0067] Therefore, the applicant of this application performs interactive reinforcement based on distinguishable physical excitation on the temperature-pressure correlation feature matrix M, which is expressed as:

[0068]

[0069] Where M is the temperature-pressure correlation feature matrix, a and b are predetermined hyperparameters, and represents positional addition and subtraction of the feature matrix, division represents dividing each position of the feature matrix by the corresponding value, and Cov(·) represents the convolution operation through a single convolutional layer, and X is the optimized temperature-pressure correlation feature matrix.

[0070] Here, the interaction reinforcement based on distinguishable physical excitation is used to enhance the interaction between the feature space and the solution space of the classification problem during the backpropagation process of gradient descent. It extracts and simulates actionable features in a manner similar to physical excitation. Thus, a general-purpose low-dimensional differentiable physical excitation method is used to obtain a physical expression of actionable features with gradient distinguishability, thereby strengthening the active parts within the temperature-pressure correlation feature matrix M during training, thereby improving the training speed of the optimized temperature-pressure correlation feature matrix X in the decoding regression task and the accuracy of the decoding results of the trained features. In this way, it is possible to accurately warn of battery pack shell fires and accurately determine the battery fire risk level.

[0071] In step S160, the optimized temperature-pressure correlation feature matrix is ​​passed through a decoder to obtain a decoded value representing the battery fire risk level. Specifically, a decoder regression is performed using the temporal coordinated correlation feature distribution information of the gas temperature and pressure to obtain the decoded value representing the battery fire risk level.

[0072] Specifically, in the embodiment of the present application, the decoder is used to perform decoding regression on the optimized temperature-pressure correlation feature matrix using the following formula to obtain the decoded value; wherein, the formula is: Where X is the optimized temperature-pressure correlation feature matrix, Y is the decoded value, and W is the weight matrix. Represents matrix multiplication.

[0073] In step S170, a warning is generated based on a comparison between the decoded value and a predetermined threshold. In one specific example of this application, a fire warning is generated in response to the decoded value being greater than the predetermined threshold. This serves to determine the battery fire risk level, allowing sufficient time to develop a self-rescue plan, avoid serious accidents caused by the fire, and protect the lives of passengers. For example, the warning can be provided by a buzzer or horn.

[0074] In summary, the vehicle fire warning method based on the embodiment of the present application is explained. It uses deep learning-based artificial intelligence detection technology to capture the dynamic time-series characteristic information of the temperature and pressure of the gas discharged from the gas exhaust part of the battery pack through a convolutional neural network model, and mines the high-dimensional time-series implicit correlation feature distribution information between the two as the input of the decoder. In this way, it can accurately warn of fires in the battery pack housing and judge the battery fire risk level at the same time, thereby reserving sufficient time to formulate corresponding self-rescue plans, avoiding serious accidents caused by fires, and protecting the lives of passengers.

[0075] Exemplary Systems

[0076] Figure 4 FIG is a block diagram of a vehicle fire warning system according to an embodiment of the present application. Figure 4As shown, according to an embodiment of the present application, the vehicle fire warning system 100 includes: a monitoring module 110, which is used to obtain temperature values ​​and pressure values ​​of the gas generated by the monitored battery pack at multiple predetermined time points within a predetermined time period; a data structuring module 120, which is used to arrange the temperature values ​​and pressure values ​​of the multiple predetermined time points into a temperature input vector and a pressure input vector according to the time dimension; an association coding module 130, which is used to association code the temperature input vector and the pressure input vector to obtain a temperature-pressure association input matrix; a feature extraction module 140, which is used to pass the temperature-pressure association input matrix through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure association feature matrix; an eigenvalue discrimination module 150, which is used to perform eigenvalue discrimination on the temperature-pressure association feature matrix to obtain an optimized temperature-pressure association feature matrix; a decoding regression module 160, which is used to pass the optimized temperature-pressure association feature matrix through a decoder to obtain a decoded value for indicating the battery fire risk level; and an early warning generation module 170, which is used to generate an early warning prompt based on a comparison between the decoded value and a predetermined threshold.

[0077] Here, those skilled in the art will appreciate that the specific functions and operations of the various units and modules in the vehicle fire warning system 100 have been described above with reference to FIG. Figures 1 to 3 The description of the vehicle fire warning method has been described in detail, and therefore, its repeated description will be omitted.

[0078] As described above, the vehicle fire warning system 100 according to an embodiment of the present application can be implemented in various terminal devices, such as a vehicle fire warning server. In one example, the vehicle fire warning system 100 according to an embodiment of the present application can be integrated into a terminal device as a software module and / or a hardware module. For example, the vehicle fire warning system 100 can be a software module in the terminal device's operating system, or it can be an application developed specifically for the terminal device. Of course, the vehicle fire warning system 100 can also be one of the terminal device's many hardware modules.

[0079] Alternatively, in another example, the vehicle fire warning system 100 and the terminal device may also be separate devices, and the vehicle fire warning system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0080] Exemplary electronic devices

[0081] Below, reference Figure 5 To describe the electronic device according to the embodiment of the present application. Figure 5 FIG is a block diagram of an electronic device according to an embodiment of the present application. Figure 5As shown, the electronic device 10 includes one or more processors 11 and a memory 12 .

[0082] The processor 11 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0083] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may run the program instructions to implement the functions of the vehicle fire warning method of each embodiment of the present application described above and / or other desired functions. Various contents such as the temperature value and pressure value of the gas generated by the monitored battery pack may also be stored in the computer-readable storage medium.

[0084] In one example, the electronic device 10 may further include an input device 13 and an output device 14 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0085] The input device 13 may include, for example, a keyboard, a mouse, and the like.

[0086] The output device 14 can output various information to the outside, including early warning prompts, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0087] Of course, to simplify, Figure 5 Only some of the components related to the present application in the electronic device 10 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device 10 may further include any other appropriate components according to specific application scenarios.

[0088] Exemplary computer program products and computer-readable storage media

[0089] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to perform the steps in the functions of the vehicle fire warning method according to various embodiments of the present application described in the above-mentioned "Exemplary Method" section of this specification.

[0090] The computer program product may be written in any combination of one or more programming languages ​​to implement the program code for performing the operations of the embodiments of the present application, including object-oriented programming languages ​​such as Java, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0091] In addition, an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enable the processor to execute the steps in the functions of the vehicle fire warning method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0092] The computer-readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0093] The basic principles of the present application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this application are merely illustrative and not restrictive, and it should not be assumed that these advantages, strengths, and effects are required of each embodiment of this application. In addition, the specific details disclosed above are merely illustrative and facilitating understanding, and are not restrictive. The above details do not limit this application to necessarily being implemented using the above specific details.

[0094] The block diagrams of the devices, devices, equipment, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "include," "comprise," "have," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0095] It should also be noted that in the apparatus, device, and method of the present application, each component or each step can be decomposed and / or recombined, and such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0096] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0097] The above description has been provided for the purpose of illustration and description. Furthermore, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A vehicle fire warning method, characterized in that: include: Obtaining temperature and pressure values ​​of gas generated by the monitored battery pack at a plurality of predetermined time points within a predetermined time period; Arranging the temperature values ​​and pressure values ​​at the plurality of predetermined time points into a temperature input vector and a pressure input vector respectively according to the time dimension; performing correlation coding on the temperature input vector and the pressure input vector to obtain a temperature-pressure correlation input matrix; Passing the temperature-pressure correlation input matrix through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix; performing eigenvalue differentiation on the temperature-pressure correlation characteristic matrix to obtain an optimized temperature-pressure correlation characteristic matrix; Passing the optimized temperature-pressure correlation feature matrix through a decoder to obtain a decoded value representing a battery fire risk level; as well as Based on the comparison between the decoded value and a predetermined threshold, an early warning prompt is generated.

2. The vehicle fire warning method according to claim 1, characterized in that: Correlation encoding is performed on the temperature input vector and the pressure input vector to obtain a temperature-pressure correlation input matrix, comprising: The temperature input vector and the pressure input vector are associated and encoded using the following formula to obtain a temperature-pressure associated input matrix; Wherein, the formula is: in represents the transposed vector of the temperature input vector, V b represents the pressure input vector, M represents the temperature-pressure correlation input matrix, Represents matrix multiplication.

3. The vehicle fire warning method according to claim 2, characterized in that: The temperature-pressure correlation input matrix is ​​passed through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix, including: Using the convolution layer of the first layer of the convolutional neural network model to convolve the input data with a first convolution kernel to generate a first convolution feature map; Performing pooling processing on the first convolutional feature map along the channel dimension using the pooling layer of the first layer of the convolutional neural network model to generate a first pooling feature matrix; Performing activation processing on the first pooling feature matrix using the activation layer of the first layer of the convolutional neural network model to generate a first activation feature map; Using a convolution layer of a second layer of the convolutional neural network model to perform convolution processing on the input data with a second convolution kernel to generate a second convolution feature map, where the second convolution kernel and the first convolution kernel are transposed of each other; Performing a pooling process on the second convolutional feature map along the channel dimension using the pooling layer of the second layer of the convolutional neural network model to generate a second pooling feature matrix; and Performing activation processing on the second pooling feature matrix using the activation layer of the second layer of the convolutional neural network model to generate a second activation feature map; Among them, the output of the last layer of the convolutional neural network model is the temperature-pressure correlation feature matrix, and the input of the first layer of the convolutional neural network model is the temperature-pressure correlation input matrix.

4. The vehicle fire warning method according to claim 3, characterized in that: Performing eigenvalue differentiation on the temperature-pressure correlation characteristic matrix to obtain an optimized temperature-pressure correlation characteristic matrix includes: The temperature-pressure correlation feature matrix is ​​interactively reinforced based on distinguishable physical excitation using the following formula to obtain the optimized temperature-pressure correlation feature matrix; Wherein, the formula is: Where M is the temperature-pressure correlation feature matrix, a and b are predetermined hyperparameters, and represents positional addition and subtraction of the feature matrix, division represents dividing each position of the feature matrix by the corresponding value, and Cov(·) represents the convolution operation through a single convolutional layer, and X is the optimized temperature-pressure correlation feature matrix.

5. The vehicle fire warning method according to claim 4, characterized in that: Passing the optimized temperature-pressure correlation feature matrix through a decoder to obtain a decoded value representing a battery fire risk level includes: The decoder is used to perform decoding regression on the optimized temperature-pressure correlation feature matrix using the following formula to obtain the decoding value; wherein the formula is: Where X is the optimized temperature-pressure correlation feature matrix, Y is the decoded value, and W is the weight matrix. Represents matrix multiplication.

6. The vehicle fire warning method according to claim 5, characterized in that: The temperature value is acquired by a temperature sensor disposed on the monitored battery pack of the vehicle, and / or the pressure value is acquired by a pressure sensor disposed on the monitored battery pack of the vehicle.

7. The vehicle fire warning method according to any one of claims 1 to 6, characterized in that: Provide early warning via buzzer or whistle.

8. A vehicle fire warning system, characterized in that: include: A monitoring module, configured to obtain temperature and pressure values ​​of gas generated by the monitored battery pack at a plurality of predetermined time points within a predetermined time period; a data structuring module, configured to arrange the temperature values ​​and pressure values ​​at the plurality of predetermined time points into a temperature input vector and a pressure input vector respectively according to a time dimension; an association coding module, configured to perform association coding on the temperature input vector and the pressure input vector to obtain a temperature-pressure association input matrix; a feature extraction module, configured to pass the temperature-pressure correlation input matrix through a convolutional neural network model using mutually transposed convolution kernels in adjacent layers to obtain a temperature-pressure correlation feature matrix; an eigenvalue distinguishing module, configured to perform eigenvalue distinguishing on the temperature-pressure correlation characteristic matrix to obtain an optimized temperature-pressure correlation characteristic matrix; a decoding regression module, configured to pass the optimized temperature-pressure correlation feature matrix through a decoder to obtain a decoded value representing a battery fire risk level; as well as The warning generation module is used to generate a warning prompt based on the comparison between the decoded value and a predetermined threshold.

9. The vehicle fire warning system according to claim 8, characterized in that: The associated coding module is further configured to: The temperature input vector and the pressure input vector are associated and encoded using the following formula to obtain a temperature-pressure associated input matrix; Wherein, the formula is: in represents the transposed vector of the temperature input vector, V b represents the pressure input vector, M represents the temperature-pressure correlation input matrix, Represents matrix multiplication.

10. The vehicle fire warning system according to claim 9, characterized in that: The characteristic value distinguishing module is further used to: The temperature-pressure correlation feature matrix is ​​interactively reinforced based on distinguishable physical excitation using the following formula to obtain the optimized temperature-pressure correlation feature matrix; Wherein, the formula is: Where M is the temperature-pressure correlation feature matrix, a and b are predetermined hyperparameters, and represents positional addition and subtraction of the feature matrix, division represents dividing each position of the feature matrix by the corresponding value, and Cov(·) represents the convolution operation through a single convolutional layer, and X is the optimized temperature-pressure correlation feature matrix.

Images