Lithium ion battery thermal runaway early warning method and system based on multi-modal reconstruction fusion

Abstract

The invention provides a lithium ion battery thermal runaway early warning method and system based on multi-modal reconstruction fusion, and relates to the technical field of lithium ion battery thermal runaway early warning, and the method comprises the steps: obtaining a sound time domain signal of a lithium ion battery, extracting a battery thermal runaway sound time domain feature and a sound frequency domain feature through time-frequency transformation and phase-space reconstruction, and carrying out the early warning of the thermal runaway of the lithium ion battery; then, time domain and frequency domain features are processed through GRU and an extrusion excitation attention mechanism respectively, and finally, sound signal recognition is achieved through cross-modal attention fusion and a linear classifier. The safety of the lithium ion battery can be effectively improved, the thermal runaway risk is warned in time, and potential safety hazards are avoided.

Description

Technical Field

[0001] The present disclosure relates to the technical field of thermal runaway warning for lithium-ion batteries, and in particular to a thermal runaway warning method and system for lithium-ion batteries based on multi-modal reconstruction and fusion. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Lithium-ion batteries, with their high energy density, long cycle life and stable charge and discharge performance, have become indispensable energy storage devices in the fields of smartphones, electric vehicles and renewable energy storage, and are of great significance to improving equipment efficiency and promoting green energy transformation. However, under extreme conditions such as overcharging, high temperature or physical damage, lithium-ion batteries are at risk of thermal runaway. This phenomenon is accompanied by a sharp rise in the temperature inside the battery, which may trigger electrolyte decomposition, gas expansion and even short circuit, ultimately leading to serious safety threats such as fire or explosion, posing a major hidden danger to personnel safety and equipment operation.

[0004] At present, the thermal runaway warning methods of lithium-ion batteries mainly rely on signals such as temperature, voltage, current and gas detection, but these methods have problems such as delayed response, high cost or complex deployment. For example, the invention patent-CN202411256611.2 couples the battery's state of charge, health state, thermodynamics and chemical reaction models to calculate the total heat generation rate and temperature change to predict thermal runaway, but this method is more complex and the adaptability to different working conditions still needs to be optimized. Invention patent-CN202410947610.6 proposes a thermal runaway warning method and system based on battery gas detection, but this method has high deployment requirements and the gas sensor is expensive. In contrast, acoustic signals have the advantages of low cost and non-contact monitoring, and show great potential in the early warning of thermal runaway of lithium-ion batteries.

[0005] However, acoustic signals are susceptible to noise interference in complex environments, and key features cannot be effectively extracted. Before traditional monitoring indicators such as voltage, current and temperature show obvious changes, it is impossible to achieve high-sensitivity early warning of thermal runaway of lithium-ion batteries. In addition, there are some high-frequency fluctuations or anomalies in acoustic signals. Existing deep learning models cannot fully capture the complex characteristics of fluctuations in the process of battery thermal runaway, and the prediction accuracy is low. Summary of the invention

[0006] In order to solve the above problems, the present disclosure proposes a lithium-ion battery thermal runaway warning method and system based on multi-modal reconstruction fusion, which uses time-frequency transformation and phase space reconstruction to extract battery thermal runaway sound signal features, and processes time domain and frequency domain features respectively through gated recurrent unit (Gated Recurrent Unit, GRU) and squeeze-and-excitation attention (Squeeze-and-Excitation, SE). Finally, fault signals are recognized through cross-modal attention fusion and linear classifier, which effectively improves the safety of lithium-ion batteries, timely warns of thermal runaway risks, and avoids potential safety hazards.

[0007] According to some embodiments, the present disclosure adopts the following technical solutions: The thermal runaway early warning method for lithium-ion batteries based on multi-modal reconstruction fusion includes: Acquire the sound time domain signal of the lithium-ion battery, and transform the sound time domain signal into the sound frequency domain signal by Fourier transform; Based on the phase space reconstruction method, the time delay and embedding dimension are determined, and the sound time domain signal and the sound frequency domain signal are respectively embedded in high dimensions to reconstruct the high-dimensional phase space, and the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space are respectively extracted to obtain the time domain features and the frequency domain features; The time domain features are input into the gated recurrent unit for time modeling, and the frequency domain features are input into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; The relationship between the weighted frequency domain features and the time domain features after time modeling is calculated through the cross-modal attention mechanism, and the weighted frequency domain features and the time domain features after time modeling are dynamically fused to obtain the fused features. The fused features are input into the fully connected layer, and the thermal runaway classification results of lithium-ion batteries are output through linear transformation and activation function.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: The lithium-ion battery thermal runaway warning system based on multi-modal reconstruction fusion includes: A signal acquisition module is used to acquire a sound time domain signal of a lithium-ion battery and convert the sound time domain signal into a sound frequency domain signal by applying Fourier transform; A feature extraction module is used to determine the time delay and embedding dimension based on the phase space reconstruction method, perform high-dimensional embedding on the sound time domain signal and the sound frequency domain signal to reconstruct the high-dimensional phase space, and extract the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space to obtain the time domain features and the frequency domain features; The feature fusion module is used to input the time domain features into the gated recurrent unit for time modeling, and input the frequency domain features into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; the relationship between the weighted frequency domain features and the time domain features after time modeling is calculated through the cross-modal attention mechanism, and the weighted frequency domain features are dynamically fused with the time domain features after time modeling to obtain fused features; The classification module is used to input the fusion features into the fully connected layer, and output the classification results of thermal runaway of lithium-ion batteries through linear transformation and activation function.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product includes a computer program, and when the computer program is executed by a processor, the lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion is implemented.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion is implemented.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device comprises: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device implements the lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion.

[0012] Compared with the prior art, the present invention has the following beneficial effects: The disclosed lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion, since there are three main sound signals in the thermal runaway of lithium-ion batteries, namely noise, thermal runaway expansion sound and quasi-expansion sound, uses the unique sound characteristics of acoustic wave information for diagnosis, and can give early and highly sensitive early warning of thermal runaway of lithium-ion batteries before traditional monitoring indicators such as voltage, current and temperature show obvious changes.

[0013] The lithium-ion battery thermal runaway warning method based on multimodal reconstruction fusion disclosed in the present invention uses time-frequency transformation technology to decompose the time-frequency signal and extract the noise and target signal, because noise usually has random and wide spectral characteristics and large frequency variations. The time-domain and frequency-domain signals are embedded in high dimensions through phase space reconstruction to capture the short-term dynamic and frequency characteristics of the signal, thereby accurately extracting the time variation pattern and spectral fluctuation characteristics in the battery sound signal, especially the unique characteristics of thermal runaway expansion sound and noise, providing a comprehensive description for effectively identifying the thermal runaway phenomenon.

[0014] The disclosed lithium-ion battery thermal runaway warning method based on multimodal reconstruction fusion, based on the sound characteristics of lithium-ion batteries, establishes a thermal runaway sound classifier model with a squeeze-excitation attention mechanism and a cross-modal attention mechanism, which is used to classify and identify noise, thermal runaway expansion sound, and noise similar to expansion sound, ensuring that thermal runaway events can trigger warnings in a timely manner and having a good thermal runaway recognition rate.

[0015] Compared with traditional methods, the disclosed lithium-ion battery thermal runaway warning method and system based on multi-modal reconstruction fusion does not rely on expensive hardware equipment configuration, thereby effectively reducing the investment cost of the warning system. In addition, the disclosed method is based on sound signal analysis, has high flexibility and adaptability, can be smoothly deployed and implemented in different application scenarios, and provides a new solution for the safety management of lithium-ion batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0017] Figure 1 This is a flowchart of the implementation of the lithium-ion battery thermal runaway warning method based on multi-modal reconstruction fusion according to an embodiment of the present disclosure; Figure 2 This is a sound waveform diagram of a 125% SOC thermal runaway experiment in an embodiment of the present disclosure; Figure 3 Schematic diagram of the sound classifier confusion matrix of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0018] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0019] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0020] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0021] Example 1 In one embodiment of the present disclosure, a lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion is provided, and the method steps are as follows: Step 1: Obtain the sound time domain signal of the lithium-ion battery, and apply Fourier transform to transform the sound time domain signal into a sound frequency domain signal; Step 2: Based on the phase space reconstruction method, determine the time delay and embedding dimension, perform high-dimensional embedding on the sound time domain signal and the sound frequency domain signal to reconstruct the high-dimensional phase space, extract the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space, and obtain the time domain features and frequency domain features; Step 3: Input the time domain features into the gated recurrent unit for time modeling, and input the frequency domain features into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; Step 4: Calculate the relationship between the weighted frequency domain features and the time domain features after time modeling through the cross-modal attention mechanism, dynamically fuse the weighted frequency domain features with the time domain features after time modeling to obtain the fused features, input the fused features into the fully connected layer, and output the thermal runaway classification results of lithium-ion batteries through linear transformation and activation function.

[0022] As an embodiment, the lithium-ion battery thermal runaway early warning method based on multimodal reconstruction fusion disclosed in the present invention, in the thermal runaway experiment of the lithium-ion battery, mainly found the existence of three sound signals: noise, thermal runaway expansion sound and noise similar to expansion sound. Noise usually has a random and wide spectrum feature, and the frequency changes greatly. Therefore, the present invention adopts time-frequency transformation technology, which can decompose the time-frequency signal and extract the noise and target signal. The thermal runaway expansion sound has a unique frequency pattern and shows obvious characteristics in the time domain and frequency domain. In order to effectively capture these characteristics, the phase space reconstruction technology is used to convert the time domain and frequency domain signals into high-dimensional signals. At the same time, the Gated Recurrent Unit (GRU) module is used for time domain feature extraction, which can process such progressive signals. Expansion-like noise has similar amplitude characteristics to expansion sound and is prone to interference. Therefore, a squeeze-and-excitation (SE) module is designed to weight the frequency domain features, enhance the frequency components of thermal runaway expansion sound, and suppress interference. Finally, the cross-modal attention mechanism fuses the time domain and frequency domain features, and a linear classifier is used to classify and identify the three sound signals, effectively distinguishing between noise, thermal runaway expansion sound, and expansion-like noise, ensuring that thermal runaway events can trigger warnings in a timely manner. The specific implementation process is as follows: Step 1: Obtain the sound time domain signal of the lithium-ion battery, and apply Fourier transform to transform the sound time domain signal into a sound frequency domain signal; Specifically, a sound sensor is installed at a set distance from the center of the lithium-ion battery to collect the sound time domain signal of the lithium-ion battery in real time.

[0023] Furthermore, the Fourier transform is applied to transform the sound time domain signal into a sound frequency domain signal. Specifically, the Fourier transform transforms the time domain signal Convert to frequency domain signal The conversion process is as follows:

[0024] in, is the frequency domain signal, is the time domain signal, is a frequency variable, It's time. is an imaginary unit. The Fourier transform performs a weighted summation of the time domain signal over the entire time interval through an integration operation to obtain the corresponding frequency components. The converted frequency domain signal contains the amplitude and phase information of each frequency component, which can reveal the distribution characteristics of the signal at different frequencies.

[0025] Step 2: Based on the phase space reconstruction method, determine the time delay and embedding dimension, perform high-dimensional embedding on the sound time domain signal and the sound frequency domain signal to reconstruct the high-dimensional phase space, extract the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space, and obtain the time domain features and frequency domain features; Specifically, based on the phase space reconstruction method, the sound time domain signal and the sound frequency domain signal are respectively embedded into the high-dimensional phase space to reconstruct the high-dimensional phase space, wherein, according to the time delay embedding theorem, the time delay and embedding dimension are determined so that the reconstructed phase space retains the structure of the original dynamic system and its evolution characteristics.

[0026] First, determine the appropriate time delay. The time delay is calculated as: (1) in, It refers to the sound time domain signal. Refers to the current moment, Refers to the sum of the current time and the delay time. is the autocorrelation value at time delay, is the mean value of the time series, N is the total length of the time series. The first zero or the first local minimum of the autocorrelation function is usually used as the time lag.

[0027] Secondly, the embedding dimension is determined. The embedding dimension in the phase space reconstruction is determined based on the pseudo nearest neighbor method. By calculating the distance between adjacent points in different dimensions, it is determined whether there are pseudo nearest neighbor points. The calculation process is: (2) in, m is the embedding dimension, It's a time delay.

[0028] If the distance between adjacent points increases significantly after increasing the dimension, the point is considered to be a pseudo neighbor. By calculating the proportion of pseudo neighbor points, the appropriate embedding dimension is selected. m , which minimizes the error and effectively reconstructs the phase space of the original dynamical system.

[0029] The final reconstructed trajectory matrix can be written as: (3) Furthermore, according to the reconstructed high-dimensional phase space, the nonlinear features of the sound time domain signal and the sound frequency domain signal are extracted from different angles. First, the phase space of the sound signal is reconstructed to obtain its intrinsic dynamic structure. Then, five nonlinear features are extracted in the high-dimensional phase space of the time domain signal and the frequency domain signal, including approximate entropy, Shannon entropy, fractal coefficient, correlation dimension and recurrence rate. Among them, the approximate entropy and Shannon entropy are extracted by the calculation method based on probability distribution, the fractal coefficient is extracted based on the box counting method or the Higuchi fractal dimension algorithm, the correlation dimension is extracted by the Grassberger-Procaccia algorithm, and the recurrence rate feature is extracted by recurrence plot analysis. For each signal, five phase space reconstruction features of the same different angles are extracted, namely, the phase space reconstruction features of the sound time domain signal are approximate entropy, Shannon entropy, fractal coefficient, correlation dimension and recurrence rate features, which are unified as time domain features; the phase space reconstruction features of the sound frequency domain signal in the high-dimensional phase space are extracted, and the phase space reconstruction features of the sound frequency domain signal are also approximate entropy, Shannon entropy, fractal coefficient, correlation dimension and recurrence rate features, which are unified as frequency domain features. As shown in Table 1, Table 1 shows the extracted features after phase space reconstruction.

[0030] Table 1 Features extracted after phase space reconstruction

[0031] Step 3: Input the time domain features into the gated recurrent unit for time modeling, and input the frequency domain features into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; Specifically, after completing the phase space reconstruction feature extraction, the present disclosure further designs a deep learning framework that combines time domain and frequency domain features. The time domain features and frequency domain features respectively reflect the dynamic response of the battery in the time dimension and the change of the frequency component. The two complement each other and jointly improve the prediction performance of the model in thermal runaway detection.

[0032] The time domain features and frequency domain features are further input into the sound classifier model for processing, which includes the processing of gated recurrent units and channel attention (squeeze excitation). First, the time domain features are input into the gated recurrent unit, which is composed of a convolutional neural network and a GRU. The time domain features are processed by the convolutional neural network (CNN) and the GRU. The CNN can capture the local time dependency in the signal, while the GRU can process the long-term dependency of the sound signal of thermal runaway of lithium-ion batteries through its dynamic time modeling capability.

[0033] Specifically, a convolutional neural network (CNN) and a gated recurrent unit (GRU) are combined to process the time domain features of the thermal runaway sound signal of lithium-ion batteries, where CNN is mainly used to extract local time dependency features, while GRU captures long-term dependencies through its dynamic time modeling capabilities. First, two one-dimensional convolutional layers (Conv1D) are used to ensure that the feature patterns in short time series can be captured using a small receptive field (kernel size is 3), while the nonlinear expression ability is enhanced through the ReLU activation function to prevent the gradient vanishing problem. In addition, the pooling layer further reduces the time series dimension, improves computational efficiency, and extracts the most significant features globally. Subsequently, a bidirectional GRU is used to simultaneously consider the dynamic change characteristics of the forward and backward signals, thereby enhancing the modeling capabilities of complex time dependencies. Finally, the time series features output by the GRU are globally averaged to form a fixed-dimensional representation, ensuring that the model can capture both local short-term features and model the long-term evolution trend of the signal during learning, thereby effectively improving the classification accuracy of the thermal runaway sound signal of lithium-ion batteries.

[0034] On the other hand, the frequency domain features are weighted by the SE module. The SE module first performs global average pooling on each feature channel, extracts global information and generates a vector describing the importance of the channel. The vector is then compressed and reconstructed through a two-layer fully connected network. The final output is used as the channel weight coefficient to weight the frequency domain features, so that the model automatically focuses on the key frequency components, as follows: (4) in, x is the frequency domain feature of the input, W 1 and W 2 is the weight matrix in the network, 𝜎 is the Sigmoid activation function, and the output y is the weighted frequency domain feature.

[0035] Step 4: Calculate the relationship between the weighted frequency domain features and the time domain features after time modeling through the cross-modal attention mechanism, dynamically fuse the weighted frequency domain features with the time domain features after time modeling to obtain the fused features, input the fused features into the fully connected layer, and output the thermal runaway classification results of lithium-ion batteries through linear transformation and activation function.

[0036] Specifically, the cross-modal attention mechanism is used to calculate the relationship between the weighted frequency domain features and the time domain features after time modeling, and the features of each modality are dynamically weighted to achieve the organic fusion of information and obtain the fusion features. The mathematical formula of the cross-modal attention mechanism is: Attention = softmax( QK T )(5) Fused_features = Attention× V (6) Among them, the time domain features are Q and K , the frequency domain features are V , Fused_features To fuse features, T refers to the transpose of the K matrix.

[0037] The above method disclosed in the present invention is based on the complementarity of time domain and frequency domain features. The time domain features effectively capture the dynamic changes of the signal and reflect the mutations and instantaneous fluctuations in the thermal runaway process, and are therefore used to direct attention to key changes in time. The frequency domain features reveal the frequency components of the signal, can identify high-frequency fluctuations and anomalies, and use them as values ​​to provide stable frequency information, thereby enhancing the weighted effect of the time domain features. Through this cross-modal fusion, the model can more comprehensively capture the complex features of the battery thermal runaway process and improve prediction accuracy.

[0038] Furthermore, the fused features are input to the fully connected layer. This layer goes through further linear transformation and activation function and finally outputs the classification result. This process is implemented by the following formula: Output =softmax[ fc ( Fused_features )](7) in, fc is a fully connected layer. The probability of each classification is output through the softmax function, and the final classification result is determined according to the maximum probability, thereby realizing the identification and early warning of thermal runaway of lithium-ion batteries. The pseudo code of the thermal runaway sound classifier is shown in Table 2.

[0039] Table 2 Thermal runaway sound classification algorithm framework

[0040] Simulation experiment The experimental platform disclosed in this paper is designed to study the thermal runaway characteristics of lithium-ion batteries. The core components include soft-pack batteries, charge and discharge test instruments (Xingyun 5V30A battery cell energy feedback system), host computers and accelerated calorimeters. In order to fully capture the information during the thermal runaway process, a sound sensor was arranged 20 cm away from the center of the battery to monitor the acoustic signal in real time, and a camera was installed on the top cover of the accelerated calorimeter to ensure data synchronization and accurate recording of experimental phenomena. This paper uses the C rate to describe the charge / discharge current of the battery, where the charging rate of data set 1 is 3.5C (14A). SOC (State of Charge) indicates the ratio of the current charge of the battery to the nominal capacity, reflects the remaining power of the battery, and plays a key role in evaluating battery performance and life. Data set 2 was collected at an SOC of 125%, indicating that the battery was in an overcharged state. Data set 3 was collected at an SOC of 110%.

[0041] Take the 125% SOC experiment as an example. Figure 2 The correlation between the sound signal and the battery status is demonstrated. By comparing the high amplitude moments with the changes in the battery status. Although the sound amplitude reached peaks at t1, t2, t3 and t4, the battery only expanded at t3, while no corresponding changes were observed at t1, t2 and t4. This shows that high amplitude is not always directly related to battery expansion, and there may be noise interference similar to thermal runaway. For these three sound signals, the sound classifier based on phase space reconstruction and hybrid attention mechanism disclosed in the present invention is used for predictive classification.

[0042] In order to verify the sound classifier of the present invention, 80% of the data obtained in the experiment was used as the training set, and the remaining 20% ​​of the data was used as the test set. In order to eliminate accidental factors, the average value of 10 experiments was used. After the model training was completed, the confusion matrix result was as follows: Figure 3 shown. Figure 3 1 represents noise, 2 represents noise similar to expansion sound, and 3 represents expansion sound. Experimental results show that the sound classification algorithm proposed in the present disclosure can effectively process the noise in battery sound classification, and the average recognition accuracy rate reaches 98.6%.

[0043] By mapping the features extracted by phase space reconstruction back to the originally collected thermal runaway sound sequence, the occurrence time of thermal runaway and the warning time based on sound signals can be obtained, as shown in Table 3. By analyzing the results of three experiments, it is found that the thermal runaway warning time based on sound signals can trigger the warning in advance about half of the occurrence time of thermal runaway.

[0044] Table 3 Thermal runaway warning time and occurrence time

[0045] In summary, the thermal runaway warning method for lithium-ion batteries based on multimodal reconstruction fusion proposed in this disclosure combines phase space reconstruction and hybrid attention mechanism. This warning method does not require a high-precision battery mathematical model or a complex feature acquisition hardware system, effectively shortening the modeling cycle and reducing hardware costs. This technology shows significant application potential and practical value in enhancing the safety of battery systems.

[0046] Example 2 In one embodiment of the present disclosure, a lithium-ion battery thermal runaway warning system based on multi-modal reconstruction fusion is provided, comprising: A signal acquisition module is used to acquire a sound time domain signal of a lithium-ion battery and convert the sound time domain signal into a sound frequency domain signal by applying Fourier transform; A feature extraction module is used to determine the time delay and embedding dimension based on the phase space reconstruction method, perform high-dimensional embedding on the sound time domain signal and the sound frequency domain signal to reconstruct the high-dimensional phase space, and extract the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space to obtain the time domain features and the frequency domain features; The feature fusion module is used to input the time domain features into the gated recurrent unit for time modeling, and input the frequency domain features into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; the relationship between the weighted frequency domain features and the time domain features after time modeling is calculated through the cross-modal attention mechanism, and the weighted frequency domain features are dynamically fused with the time domain features after time modeling to obtain fused features; The classification module is used to input the fusion features into the fully connected layer, and output the classification results of thermal runaway of lithium-ion batteries through linear transformation and activation function.

[0047] Example 3 In one embodiment of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the method for early warning of thermal runaway of a lithium-ion battery based on multi-modal reconstruction fusion is implemented.

[0048] Example 4 In one embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the lithium-ion battery thermal runaway warning method based on multi-modal reconstruction fusion is implemented.

[0049] Example 5 In one embodiment of the present disclosure, an electronic device is provided, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the lithium-ion battery thermal runaway warning method based on multi-modal reconstruction fusion.

[0050] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0051] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0052] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion, characterized in that: include: Acquire the sound time domain signal of the lithium-ion battery, and transform the sound time domain signal into the sound frequency domain signal by Fourier transform; Based on the phase space reconstruction method, the time delay and embedding dimension are determined, and the sound time domain signal and the sound frequency domain signal are respectively embedded in high dimensions to reconstruct the high-dimensional phase space, and the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space are respectively extracted to obtain the time domain features and the frequency domain features; The time domain features are input into the gated recurrent unit for time modeling, and the frequency domain features are input into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; The relationship between the weighted frequency domain features and the time domain features after time modeling is calculated through the cross-modal attention mechanism, and the weighted frequency domain features and the time domain features after time modeling are dynamically fused to obtain the fused features. The fused features are input into the fully connected layer, and the thermal runaway classification results of lithium-ion batteries are output through linear transformation and activation function.

2. The lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion according to claim 1, characterized in that: Based on the phase space reconstruction method, the sound time domain signal and the sound frequency domain signal are embedded into the high-dimensional phase space respectively. According to the time delay embedding theorem, the time delay and embedding dimension are determined so that the reconstructed phase space retains the structure and evolution characteristics of the original dynamic system. The calculation process of time delay is: in, It refers to the sound time domain signal. Refers to the current moment, Refers to the sum of the current time and the delay time. is the autocorrelation value at time delay, is the mean value of the time series, N is the total length of the time series; the first zero or the first local minimum of the autocorrelation function is used as the time lag.

3. The lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion according to claim 1, characterized in that: The embedding dimension in phase space reconstruction is determined based on the pseudo-nearest neighbor method. By calculating the distance between adjacent points in different dimensions, it is determined whether there are pseudo-nearest neighbor points. The calculation process is as follows: in, m is the embedding dimension, It's a time delay.

4. The lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion according to claim 1, characterized in that: The phase space reconstruction features of the sound time domain signal and the acoustic frequency domain signal in the high-dimensional phase space are extracted respectively, including: extracting the phase space reconstruction features of the sound time domain signal in the high-dimensional phase space, the phase space reconstruction features of the sound time domain signal are approximate entropy, Shannon entropy, fractal coefficient, correlation dimension and recurrence rate features, which are unified as time domain features; extracting the phase space reconstruction features of the acoustic frequency domain signal in the high-dimensional phase space, the phase space reconstruction features of the acoustic frequency domain signal are also approximate entropy, Shannon entropy, fractal coefficient, correlation dimension and recurrence rate features, which are unified as frequency domain features.

5. The lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion according to claim 1, characterized in that: The time domain features are input into the gated recurrent unit for time modeling, and the frequency domain features are input into the squeeze-excited attention module for weighted processing to obtain weighted frequency domain features, including: inputting the time domain features into the gated recurrent unit, capturing the local time dependencies in the time domain features through CNN, and then capturing the long-term dependencies of the time domain features through the dynamic time modeling capability of GRU; the frequency domain features are input into the squeeze-excited attention module, firstly, global average pooling is performed on each feature channel in the frequency domain, global information is extracted and a vector describing the importance of the channel is generated, the vector is compressed and reconstructed to obtain the channel weight coefficient, and the frequency domain features are weighted using the weight coefficient.

6. The lithium-ion battery thermal runaway early warning method based on multi-modal reconstruction fusion according to claim 1, characterized in that: The cross-modal attention mechanism calculates the relationship between the weighted frequency domain features and the time domain features after time modeling, and dynamically weights the features in the frequency domain and time domain modes to achieve information fusion. The mathematical formula of the cross-modal attention mechanism is: Attention = softmax( QK T ) Fused_features = Attention× V Among them, the time domain features are Q and K , the frequency domain features are V , T is the transpose of matrix K.

7. A lithium-ion battery thermal runaway warning system based on multi-modal reconstruction fusion, characterized in that: include: A signal acquisition module is used to acquire a sound time domain signal of a lithium-ion battery and convert the sound time domain signal into a sound frequency domain signal by applying Fourier transform; A feature extraction module is used to determine the time delay and embedding dimension based on the phase space reconstruction method, perform high-dimensional embedding on the sound time domain signal and the sound frequency domain signal to reconstruct the high-dimensional phase space, and extract the phase space reconstruction features of the sound time domain signal and the sound frequency domain signal in the high-dimensional phase space to obtain the time domain features and the frequency domain features; The feature fusion module is used to input the time domain features into the gated recurrent unit for time modeling, and input the frequency domain features into the squeeze-excitation attention module for weighted processing to obtain weighted frequency domain features; the relationship between the weighted frequency domain features and the time domain features after time modeling is calculated through the cross-modal attention mechanism, and the weighted frequency domain features are dynamically fused with the time domain features after time modeling to obtain fused features; The classification module is used to input the fusion features into the fully connected layer, and output the classification results of thermal runaway of lithium-ion batteries through linear transformation and activation function.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the lithium-ion battery thermal runaway warning method based on multi-modal reconstruction fusion according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by the processor, the lithium-ion battery thermal runaway warning method based on multi-modal reconstruction fusion as described in any one of claims 1 to 6 is implemented.

10. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the lithium-ion battery thermal runaway warning method based on multi-modal reconstruction fusion as described in any one of claims 1 to 6.

Images