Lithium battery pack thermal runaway prediction method, device, equipment and medium

Abstract

The invention provides a lithium battery pack thermal runaway prediction method, device and equipment and a medium, and belongs to the technical field of battery application. The method comprises the following steps: acquiring a lithium battery pack thermal imaging image, an electrical parameter and a cabin gas parameter; recognizing the thermal imaging image to obtain features, and inputting the features into a pre-constructed early warning classification model to determine a thermal runaway category; wherein the thermal runaway category comprises a thermal runaway stage; determining the state of the lithium battery pack based on the electrical and gas parameters; if the thermal runaway category and state are both in a thermal runaway stage, identifying the single battery with the highest temperature by using a target detection algorithm; and if one of the two is in the thermal runaway stage, returning to the step of inputting the image features into the early warning classification model to determine the thermal runaway category. Through comprehensive thermal imaging analysis and multi-parameter monitoring, accurate prediction, real-time monitoring and quick response of the state of the lithium battery pack are realized, so that the safety management level of the lithium battery is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of battery application technology, and in particular to a method for predicting thermal runaway of a lithium battery pack, a device for predicting thermal runaway of a lithium battery pack, an electronic device and a computer-readable storage medium. Background Art

[0002] Lithium-ion batteries are one of the most widely used energy storage devices and an important pillar of future energy management systems. Among many energy storage technologies, lithium-ion batteries have higher power density than other batteries. The safe, reliable and long-lasting use of lithium-ion batteries is becoming increasingly important. However, the explosion of energy storage power stations that occurs from time to time has forced us to re-evaluate the necessity of lithium-ion battery safety.

[0003] There are many reasons for thermal runaway of lithium-ion batteries, such as battery manufacturing defects, poor (electrical or mechanical) battery design, external (thermal, mechanical or electrical) abuse of the battery, fire in charging equipment or protection equipment, etc. Generally speaking, thermal runaway of lithium-ion batteries is divided into three stages: self-heating stage, thermal runaway stage, and thermal runaway termination stage.

[0004] At present, the main technical means for predicting thermal runaway of lithium battery packs are: (1) monitoring electrical parameters, inferring internal thermal reactions through real-time detection of abnormal fluctuations in voltage, current and other data, and judging whether thermal runaway is likely to occur; (2) arranging gas sensors to detect changes in gas composition and concentration in the cabin, and using characteristic gases to assist in judging thermal runaway conditions; (3) using thermal imaging equipment to obtain the surface temperature distribution image of the battery pack to observe whether there are signs of thermal runaway such as local overheating.

[0005] However, the existing thermal runaway prediction technology for lithium battery packs has limitations due to the fact that it only relies on electrical parameter monitoring, gas sensor detection of gas parameters, and the simple use of thermal imaging technology to predict the thermal runaway of lithium batteries. In other words, it is impossible to comprehensively judge multiple factors and dynamically monitor the thermal runaway status of lithium battery packs in real time. Summary of the invention

[0006] The purpose of the embodiments of the present invention is to provide a method, device, equipment and medium for predicting thermal runaway of a lithium battery pack to solve the above-mentioned problems.

[0007] In order to achieve the above object, an embodiment of the present invention provides a method for predicting thermal runaway of a lithium battery pack, comprising: S100, obtaining thermal imaging images, electrical parameters and cabin gas parameters of the lithium battery pack; S200, performing image recognition on the thermal imaging image of the lithium battery pack to obtain features of the thermal imaging image of the lithium battery pack; S300, inputting the features of the thermal imaging image of the lithium battery pack into a lithium battery early warning classification model to obtain a thermal runaway category of the lithium battery pack; wherein the thermal runaway category includes: a thermal runaway stage; S400, if the thermal runaway category of the lithium battery pack is the thermal runaway stage or based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, it is determined that the state of the lithium battery pack is the thermal runaway stage, then return to step S300; S500, if the thermal runaway category of the lithium battery pack is the thermal runaway stage and based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, it is determined that the state of the lithium battery pack is the thermal runaway stage, then a target detection algorithm is used to perform target detection and recognition on the thermal imaging image of the lithium battery pack to determine the single cell with the highest temperature in the lithium battery pack.

[0008] Optionally, the thermal imaging image of the lithium battery pack is obtained by: Get the temperature of the lithium battery pack; If the temperature of the lithium battery pack is lower than the first preset temperature, thermal imaging images of the lithium battery pack are intercepted at intervals of a first preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and one frame of thermal imaging images is randomly selected from the multiple frames of thermal imaging images of the lithium battery pack; If the temperature of the lithium battery pack is greater than the first preset temperature and less than the second preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the second preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting one frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; If the temperature of the lithium battery pack is greater than the second preset temperature and less than the third preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the third preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting one frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; Among them, the first preset temperature is less than the second preset temperature and less than the third preset temperature; the first preset time is less than the second preset time and less than the third preset time.

[0009] Optionally, before the step of performing image recognition on the thermal imaging image of the lithium battery pack to obtain features of the thermal imaging image of the lithium battery pack, the method further includes: The thermal imaging image of the lithium battery pack is preprocessed; wherein the preprocessing includes: normalization processing, image enhancement processing, Gaussian filtering processing and histogram equalization processing.

[0010] Optionally, performing image recognition on multiple frames of thermal imaging images of the lithium battery pack to obtain features of the thermal imaging images of the lithium battery pack includes: Determine the temperature value corresponding to each pixel of the thermal imaging image of the lithium battery pack; Determine pixels in the thermal imaging image of the lithium battery pack whose temperature value is greater than a preset temperature as marked pixels of each frame of the thermal imaging image, and determine the number of marked pixels; The total number of pixels of the thermal imaging image of the lithium battery pack is determined, and the ratio of the number of marked pixels to the total number of pixels is calculated to obtain the hot spot area ratio of the thermal imaging image of the lithium battery pack.

[0011] Optionally, the lithium battery warning classification model is trained in the following way: Obtain the hot spot area ratio of thermal imaging images of lithium battery packs of different thermal runaway categories; thermal runaway categories include: self-heating stage, thermal runaway stage and thermal runaway termination stage; The pre-built convolutional neural network model is trained using the hot spot area ratio of thermal imaging images of lithium battery packs of different thermal runaway categories. After the training is completed, a lithium battery early warning classification model is obtained.

[0012] Optionally, the electrical parameters of the lithium battery pack include: voltage and current; the in-cabin gas parameters of the lithium battery pack include: in-cabin hydrogen concentration and in-cabin carbon monoxide concentration; Based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, determine the status of the lithium battery pack, including: If any one of the voltage, current, hydrogen concentration in the cabin, and carbon monoxide concentration of the lithium battery pack is greater than the corresponding preset value, it is determined that the state of the lithium battery pack is in the thermal runaway stage; If the voltage, current, cabin hydrogen concentration, and cabin carbon monoxide concentration of the lithium battery pack are all less than corresponding preset values, it is determined that the lithium battery pack is in a safe stage.

[0013] In a second aspect of an embodiment of the present invention, a device for predicting thermal runaway of a lithium battery pack is provided, comprising: A data acquisition module, used to obtain thermal imaging images, electrical parameters and cabin gas parameters of the lithium battery pack; An image recognition module is used to perform image recognition on the thermal imaging image of the lithium battery pack to obtain the characteristics of the thermal imaging image of the lithium battery pack; The classification and recognition module is used to input the features of the thermal imaging image of the lithium battery pack into the lithium battery early warning classification model to obtain the thermal runaway category of the lithium battery pack; wherein the thermal runaway category includes: thermal runaway stage; A loop calling module, used for returning to call the classification identification module when the thermal runaway category of the lithium battery pack is the thermal runaway stage or when it is determined that the state of the lithium battery pack is the thermal runaway stage based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin; The target recognition module is used to perform target detection and recognition on the thermal imaging image of the lithium battery pack using a target detection algorithm when the thermal runaway category of the lithium battery pack is the thermal runaway stage and based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, the state of the lithium battery pack is determined to be the thermal runaway stage, and determine the single cell with the highest temperature in the lithium battery pack.

[0014] Optionally, the image recognition module is specifically used for: Determine the temperature value corresponding to each pixel of the thermal imaging image of the lithium battery pack; Determine pixels in the thermal imaging image of the lithium battery pack whose temperature value is greater than a preset temperature as marked pixels of each frame of the thermal imaging image, and determine the number of marked pixels; The total number of pixels of the thermal imaging image of the lithium battery pack is determined, and the ratio of the number of marked pixels to the total number of pixels is calculated to obtain the hot spot area ratio of the thermal imaging image of the lithium battery pack.

[0015] In a third aspect of the embodiments of the present invention, an electronic device is provided, including: a processor and a memory, wherein the memory stores machine-readable instructions executable by the processor, and when the machine-readable instructions are executed by the processor, the above-mentioned method for predicting thermal runaway of a lithium battery pack is executed.

[0016] In a fourth aspect of the embodiments of the present invention, a computer-readable storage medium is provided, which stores computer instructions. When the computer instructions are executed on a computer, the computer executes the above-mentioned method for predicting thermal runaway of a lithium battery pack.

[0017] Beneficial effects of the present invention: (1) Accurate prediction: By comprehensively using the feature analysis of thermal imaging images and the monitoring of electrical parameters and cabin gas parameters to judge the status of the lithium battery pack, it can be relatively accurately determined whether the lithium battery pack is in the self-heating stage, thermal runaway stage or thermal runaway termination stage. When it is determined that the lithium battery pack is indeed in the thermal runaway stage (when the thermal runaway category and status meet the conditions), the target detection algorithm is used to perform target detection and identification on the thermal imaging image, which can accurately determine the single cell with the highest temperature in the lithium battery pack.

[0018] (2) Real-time dynamic monitoring: During the entire process, when certain thermal runaway categories or status conditions are met, the relevant steps will be repeated for continuous monitoring until the critical situation is accurately determined, ensuring the real-time status monitoring of the lithium battery pack and being able to capture thermal runaway-related changes in a timely manner.

[0019] (3) Improve response speed: By combining multiple parameters and characteristics for comprehensive judgment, compared with the traditional method of judging thermal runaway by relying on a single parameter (such as voltage or temperature alone), it can detect the trend of thermal runaway in lithium battery packs earlier and issue an early warning.

[0020] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings: Figure 1 is a schematic flow chart of a method for predicting thermal runaway of a lithium battery pack provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the structure of a thermal runaway prediction device for a lithium battery pack provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0024] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0025] Embodiment 1 Please refer to Figure 1 , Figure 1 : is a flow chart of a method for predicting thermal runaway of a lithium battery pack provided by an embodiment of the present invention, the method comprising: S100, obtaining thermal imaging images, electrical parameters and cabin gas parameters of the lithium battery pack; In one embodiment, the thermal imaging image of the lithium battery pack is obtained by: S110, obtaining the temperature of the lithium battery pack; S120, if the temperature of the lithium battery pack is lower than the first preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the first preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting one frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; S130, if the temperature of the lithium battery pack is greater than the first preset temperature and less than the second preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the second preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting a frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; S140, if the temperature of the lithium battery pack is greater than the second preset temperature and less than the third preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the third preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting a frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; Among them, the first preset temperature is less than the second preset temperature and less than the third preset temperature; the first preset time is less than the second preset time and less than the third preset time.

[0026] In this embodiment, the thermal imaging video is captured by varying the duration, with a long period when the temperature is low and a short period when the temperature is high, thereby effectively avoiding the waste of computing resources.

[0027] In one embodiment, the electrical parameters of the lithium battery pack include: the voltage of the lithium battery pack and the current of the lithium battery pack.

[0028] In one embodiment, the gas parameters in the cabin of the lithium battery pack include: the concentration of hydrogen in the cabin of the lithium battery pack and the concentration of carbon monoxide in the cabin of the lithium battery pack.

[0029] S200, performing image recognition on the thermal imaging image of the lithium battery pack to obtain features of the thermal imaging image of the lithium battery pack; In one embodiment, step S200 specifically includes: S210, determining a temperature value corresponding to each pixel point of the thermal imaging image of the lithium battery pack; Each pixel in the thermal imaging image has a corresponding temperature value, and a preset temperature is set as the judgment standard. This preset temperature is usually determined based on the safe temperature range of the lithium battery when it is working normally or related experience and standards. When the temperature value of a pixel is higher than the preset temperature, it means that there may be an abnormal temperature at the location of the pixel, so such pixels are marked and called marked pixels in each frame of the thermal imaging image.

[0030] S220, determining pixels in the thermal imaging image of the lithium battery pack whose temperature values ​​are greater than a preset temperature as marked pixels for each frame of the thermal imaging image, and determining the number of marked pixels; For example: the preset temperature is set to 50°C. In the thermal imaging image, there are a total of 100×100=10,000 pixels. After comparing the temperature value of each pixel one by one, it is found that the temperature values ​​of 200 pixels are greater than 50°C. Then these 200 pixels are the marked pixels of this frame of image.

[0031] S230, determining the total number of pixels of the thermal imaging image of the lithium battery pack, and calculating the ratio of the number of marked pixels to the total number of pixels to obtain a hot spot area ratio of the thermal imaging image of the lithium battery pack.

[0032] To know the size of the area covered by the thermal imaging image of the entire lithium battery pack, it is necessary to determine the total number of pixels in the image, which can be obtained through information such as the resolution of the image (for example, the number of horizontal pixels in the image multiplied by the number of vertical pixels). Then divide the number of final marked pixels by the total number of pixels, and the resulting ratio represents the area ratio of the hot spot (that is, the temperature abnormality area) on the thermal imaging image. This hot spot area ratio can intuitively reflect the relative size of the temperature abnormality area of ​​the lithium battery pack, thereby helping to determine whether the lithium battery pack has potential safety hazards such as overheating, as well as the severity of the hazard.

[0033] For example: Assuming that the total number of pixels in the thermal imaging image is 10,000 (for example, the image resolution is 100×100), and the final number of marked pixels determined in the previous steps is 100, then the hot spot area ratio is 100÷10000=0.01 (i.e. 1%), which means that 1% of the area in the thermal imaging image shows abnormal temperature, which can be used to further analyze the status of the lithium battery pack.

[0034] In one embodiment, the thermal imaging image of the lithium battery pack is preprocessed; wherein the preprocessing includes: normalization processing, image enhancement processing, Gaussian filtering processing and histogram equalization processing.

[0035] Normalization refers to scaling data according to certain rules so that it falls within a certain range, usually [0, 1] or [-1, 1]. In image processing, normalization can make each pixel value in the image scaled according to the same ratio. The advantage of this is that it can accelerate the convergence speed of subsequent processing algorithms and improve the stability of the model.

[0036] Image enhancement refers to improving the quality of an image through a series of algorithms to make the image more suitable for a specific application. Image enhancement technology can be used to improve the clarity, contrast, brightness, etc. of an image, making the features of the image more obvious and easier for human eyes to observe or machine analysis. Common image enhancement methods include histogram equalization, sharpening, denoising, adjusting brightness and contrast, etc.

[0037] Gaussian filtering is a linear smoothing filter that uses a Gaussian function as a convolution kernel to remove noise from an image while maintaining image edge information. The Gaussian filter performs a weighted average on each pixel of the image and the pixels in its neighborhood. The weight is calculated based on the Gaussian distribution (i.e., normal distribution), so the farther the pixel is from the center point, the smaller the weight.

[0038] Histogram equalization is a method to improve image contrast. It changes the grayscale distribution of each pixel in the image by changing the image histogram, making the image histogram as flat as possible, that is, the frequency of occurrence of each grayscale value is as consistent as possible. This method is particularly suitable for situations where the overall image contrast is low. Through equalization processing, the details of the image can be clearer and the visual effect can be better.

[0039] In this embodiment, the thermal imaging images after preprocessing provide a high-quality data basis for further analysis such as calculating the proportion of hot spot area, judging the degree of temperature abnormality, and evaluating the overall thermal state of the lithium battery pack. Based on better quality and more accurate images, more reasonable battery pack operation and maintenance strategies can be formulated, such as accurately determining whether the battery pack needs to be repaired, how to adjust the heat dissipation and charging and discharging strategies, etc., so as to better ensure the safety and stable operation of the lithium battery pack and extend its service life.

[0040] S300, inputting the features of the thermal imaging image of the lithium battery pack into a lithium battery early warning classification model to obtain a thermal runaway category of the lithium battery pack; wherein the thermal runaway category includes: a thermal runaway stage; The temperature change of the battery during the charging and discharging process mainly comes from the heat of chemical reactions, polarization reactions, ohmic impedance and side reaction heat generated by electrolyte decomposition during the charging and discharging process. The abnormal heat generation of battery thermal runaway mainly comes from side reaction heat. The analysis of temperature and temperature rise rate is an important judgment process and detection method in the process of lithium-ion battery thermal runaway detection. Due to various reasons such as overcharging, overdischarging, and extrusion, the temperature inside the battery cell continues to rise, and the SEI film is heated and melted until the positive and negative electrode reactions of the battery are completed, and heat is continuously generated.

[0041] Generally speaking, thermal runaway is divided into three stages: self-heating stage (50℃-140℃), thermal runaway stage (140℃-850℃), and thermal runaway termination stage (850℃-normal temperature).

[0042] The self-heating stage, also called the heat accumulation stage, begins with the dissolution of the SEI film. When the temperature of the SEI film reaches about 90°C, its dissolution phenomenon can be clearly observed. The dissolution of the SEI film exposes the negative electrode and the lithium-embedded carbon components contained in the negative electrode directly to the electrolyte. The lithium-embedded carbon reacts with the electrolyte exothermically, causing the temperature to rise. The temperature rise in turn promotes the further decomposition of the SEI film. If there is no external cooling method, this process will roll forward until the SEI film is completely decomposed.

[0043] In the thermal runaway stage, after the temperature exceeds 140°C, both the positive and negative electrode materials join the ranks of the electrochemical reaction, and the increase in the mass of the reactants makes the temperature rise faster. The parameter change that can be observed externally is a sharp drop in voltage, and the process is described as follows: after reaching this temperature range, the diaphragm begins to melt in large quantities, and the positive and negative electrodes are directly connected, causing a large-scale short circuit. At this point, thermal runaway has begun and will not stop. In a short period of time, the violent reaction generates a large amount of gas and a large amount of heat, which heats the gas. The expanding gas breaks through the battery cell shell, and phenomena such as material ejection occur. The scattered material also takes away some heat. Thermal runaway reaches its most intense state. The highest temperature is also reached at this stage.

[0044] The thermal runaway termination stage can only be terminated when all reactants are burned out.

[0045] In one embodiment, the pre-built lithium battery warning classification model is trained in the following manner: Step 1: Obtain the hot spot area ratio of the thermal imaging image of the lithium battery pack of different thermal runaway categories; wherein the thermal runaway categories include: self-heating stage, thermal runaway stage and thermal runaway termination stage; Acquire thermal imaging images: Use thermal imaging cameras to collect thermal imaging images of lithium battery packs in different states. These states may include self-heating stage, thermal runaway stage, and thermal runaway termination stage.

[0046] Calculation of hot spot area ratio: Process the collected thermal imaging images, extract the hot spots (i.e. the areas with higher temperatures in the image) and calculate their ratio to the entire image area. This ratio is an important indicator for judging the thermal state of lithium batteries.

[0047] Data annotation: The calculated hot spot area percentage data is associated with the corresponding thermal runaway category to form an annotated data set. The thermal runaway category may be normal, warning, mild thermal runaway, severe thermal runaway, etc.

[0048] Step 2: Use the hot spot area ratio of thermal imaging images of lithium battery packs of different thermal runaway categories to train the pre-built convolutional neural network model, and after the training is completed, obtain the pre-built lithium battery warning classification model.

[0049] Build a convolutional neural network (CNN) model: Design a convolutional neural network model suitable for processing image data. Convolutional neural networks can effectively extract features from images and are suitable for image classification tasks.

[0050] Convolutional Neural Network (CNN) is an efficient classification method developed in recent years. Due to its powerful automatic feature capture ability, it is widely used in electrical field image classification. Usually, CNN is mainly composed of input layer, convolution layer, pooling layer and fully connected (FC) layer. Local connection and parameter sharing in CNN reduce the number of parameters, greatly reduce training complexity and reduce overfitting. At the same time, its weight sharing also makes CNN tolerant to translation, while downsampling in the pooling layer further reduces the output parameters, allowing the model to tolerate slight deformations, thereby improving the generalization ability of the model. In CNN, in each feature extraction layer, the feature map performs convolution calculations on multiple convolution kernels, and the feature extraction layers are connected through bias calculations, activation functions and pooling operations.

[0051] Data input: The preprocessed thermal imaging image data and its hot spot area ratio are fed to the CNN model as input data.

[0052] Model training: Forward propagation: The model receives input data, extracts features through a series of convolutional layers, pooling layers, and finally outputs prediction results through a fully connected layer.

[0053] Loss function calculation: Compare the model's prediction results with the true labels, calculate the loss function value, and measure the error of the prediction.

[0054] Backpropagation: Based on the loss function value, the model weights are updated through optimization algorithms such as gradient descent to reduce the prediction error.

[0055] Model validation and optimization: Use the validation set to evaluate the model during training and monitor the generalization ability of the model.

[0056] Adjust model parameters such as learning rate, batch size, etc., as well as model structure such as the number of layers and neurons based on the verification results.

[0057] Obtaining a warning classification model: When the performance of the model on the validation set reaches the preset standard, the training is stopped. The model at this time is the pre-built lithium battery warning classification model.

[0058] S400, if the thermal runaway category of the lithium battery pack is the thermal runaway stage or based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, it is determined that the state of the lithium battery pack is the thermal runaway stage, then return to step S300; It is understandable that when it is detected that the thermal runaway category of the lithium battery pack is the thermal runaway stage or its state is the thermal runaway stage, it returns to step S300, which builds a process monitoring and feedback adjustment mechanism, which can prevent and control the risks brought by thermal runaway, and at the same time ensure that subsequent related operations can be effectively carried out when the lithium battery pack is in a suitable state.

[0059] S500, if the thermal runaway category of the lithium battery pack is the thermal runaway stage and based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, it is determined that the state of the lithium battery pack is the thermal runaway stage, then using a target detection algorithm to perform target detection and recognition on the thermal imaging image of the lithium battery pack to determine the single cell with the highest temperature in the lithium battery pack; In one embodiment, the target detection algorithm may be a YOLO algorithm or other target detection algorithms, which is not specifically limited in the embodiment of the present invention.

[0060] YOLO (You Only Look Once) is a real-time target detection system that predicts the target bounding box and category through a single neural network, solving the regression problem. This application uses temperature data, that is, the corresponding temperature data image obtained by intercepting thermal imaging video, as the input of the YOLO network to predict the location and category of the highest temperature.

[0061] The following is an example using YOLO: (1) Data preparation stage Collecting Thermal Image Data: Collect thermal imaging image samples of lithium battery packs in different states (such as normal operation, different charging and discharging stages, different ambient temperatures, etc.), and ensure that the images cover various possible temperature distribution conditions and the appearance of single batteries in different postures and angles in the image, so that the model can learn rich features and improve its generalization ability. For example, thermal imaging of lithium battery packs can be performed in the laboratory to simulate different working conditions, or corresponding images can be collected from actual application scenarios (such as electric vehicle battery packs, energy storage power station battery packs, etc.).

[0062] Annotated image data: The collected thermal imaging images are annotated, and the position of each single battery in the image is marked with a bounding box, and its corresponding category is marked (here it can be uniformly marked as the "single battery" category). At the same time, the temperature value information corresponding to each single battery is recorded (which can be obtained through the temperature data of the thermal imaging image or other supporting temperature measurement methods) to subsequently determine the single battery with the highest temperature. The labeling work can be completed with the help of some professional image labeling tools, such as LabelImg.

[0063] (2) Model selection and configuration stage Choose the appropriate YOLO version: Choose the appropriate YOLO version based on the accuracy and speed requirements of the actual application scenario. For example, if it is in a real-time monitoring system and the accuracy requirements are not extremely stringent, a relatively lightweight and fast version such as the s version of YOLOv5 can be considered; if it is to be analyzed in a laboratory environment with high accuracy requirements, a more accurate version such as YOLOv7 may be more suitable. Different versions of YOLO differ in network structure complexity, number of parameters, and detection performance. The most suitable version can be determined through experimental comparison.

[0064] Adjust model parameters (optional): According to the characteristics of the collected thermal imaging image data, some initial parameters of the model are appropriately adjusted, such as the size of the input image (to match the actual resolution of the thermal imaging image, etc.), the size and proportion of the anchor boxes (which can be predefined according to the approximate shape and size distribution of the single battery in the image, which helps to improve the detection accuracy of the single battery target), etc., so that the model can better adapt to the data characteristics in the early stage of training, improve the training effect and detection accuracy.

[0065] (3) Model training phase Divide the dataset: The labeled thermal imaging image dataset is divided into training set, validation set and test set according to a certain ratio (such as the common 8:2 or 7:3, etc.). The training set is used for model parameter learning and optimization; the validation set is used to evaluate the performance of the model during the training process, help adjust hyperparameters, prevent overfitting, etc.; the test set is used to finally test the detection effect of the trained model on unseen data and obtain objective performance evaluation indicators.

[0066] Choose the right training environment and hardware: According to the selected YOLO version and data scale, select the appropriate deep learning training framework (such as PyTorch, TensorFlow, etc.) and the corresponding hardware platform (such as GPU server, etc.; due to the large amount of computational complexity of deep learning model training, the parallel computing capability of GPU can greatly speed up the training speed), build a training environment, and ensure that the training process can be carried out efficiently and stably.

[0067] Start training: Using the divided training set, the model training process is started in the selected training environment. By defining appropriate loss functions (such as common mean square error, cross entropy, etc., which are used to measure the difference between the prediction results and the labeled results) and optimizers (such as stochastic gradient descent, Adam, etc., which are used to update the model parameters to minimize the loss function value), the model is allowed to continuously learn the characteristics of single cells in thermal imaging images and the corresponding position and category information. After multiple training cycles (epochs), the performance of the model is gradually optimized until good performance indicators such as detection accuracy are achieved on the validation set.

[0068] (4) Target detection and determination of the single cell with the highest temperature To detect new thermal imaging images: After the model is trained, the thermal imaging image of the lithium battery pack to be analyzed is input into the model. The model will perform a forward propagation and output the detection results of each single battery in the image, including its bounding box position and corresponding category information, so as to determine the area where each single battery is located in the image.

[0069] Determine the cell with the highest temperature: According to the area where the single battery is located detected by the model, combined with the temperature value of each pixel recorded in the thermal imaging image itself (this information can be retained when the image was previously acquired), the temperature values ​​of the pixels in each detected single battery area are counted, and the temperature conditions of each single battery are compared by calculating the average temperature or finding the temperature corresponding to the highest temperature pixel in the area, etc., and finally the single battery with the highest temperature is determined.

[0070] In one embodiment, the steps of determining that the state of the lithium battery pack is in the thermal runaway stage based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin in step S400 and step S500 specifically include: First case: If any one of the voltage, current, hydrogen and carbon monoxide of the lithium battery pack is greater than the corresponding preset value, it is determined that the state of the lithium battery pack is in the thermal runaway stage; For ease of understanding, the following example is given: Assume that the preset values ​​of the parameters of the lithium battery pack we set are as follows: the voltage preset value is 4.2 volts; the current preset value is 5 amperes; the hydrogen concentration preset value is 100 ppm (parts per million concentration); the carbon monoxide concentration preset value is 80 ppm.

[0071] Now there is a lithium battery pack, and its actual conditions are detected as follows: the voltage reaches 4.5 volts, which is greater than the preset 4.2 volts; the current is 6 amperes, which is greater than the preset 5 amperes; the hydrogen concentration is 120 ppm, which is greater than the preset 100 ppm; the carbon monoxide concentration is 90 ppm, which is greater than the preset 80 ppm.

[0072] In this example, if any one of the four parameters of the lithium battery pack, namely, voltage, current, hydrogen concentration, and carbon monoxide concentration (actually all four parameters here) is greater than the corresponding preset value, it can be determined that the lithium battery pack is in the thermal runaway stage.

[0073] Second case: If the voltage, current, hydrogen and carbon monoxide of the lithium battery pack are all less than the corresponding preset values, it is determined that the state of the lithium battery pack is in a safe stage.

[0074] For ease of understanding, the following example is given: Assume that the preset values ​​of the parameters of the lithium battery pack we set are as follows: the voltage preset value is 4.2 volts; the current preset value is 5 amperes; the hydrogen concentration preset value is 100 ppm; and the carbon monoxide concentration preset value is 80 ppm.

[0075] Now there is a lithium battery pack, and its actual conditions are detected as follows: the voltage is 4.1 volts, which is less than the preset 4.2 volts; the current is 4 amperes, which is less than the preset 5 amperes; the hydrogen concentration is 80 ppm, which is less than the preset 100 ppm; the carbon monoxide concentration is 60 ppm, which is less than the preset 80 ppm.

[0076] It can be seen that the four parameters of the voltage, current, hydrogen concentration and carbon monoxide concentration of this lithium battery pack are all less than the corresponding preset values, which indicates that the lithium battery pack is in a safe state.

[0077] Beneficial effects of the present invention: (1) Accurate prediction: By comprehensively using the feature analysis of thermal imaging images and the monitoring of electrical parameters and cabin gas parameters to judge the status of the lithium battery pack, it can be relatively accurately determined whether the lithium battery pack is in the self-heating stage, thermal runaway stage or thermal runaway termination stage. When it is determined that the lithium battery pack is indeed in the thermal runaway stage (when the thermal runaway category and status meet the conditions), the target detection algorithm is used to perform target detection and identification on the thermal imaging image, which can accurately determine the single cell with the highest temperature in the lithium battery pack.

[0078] (2) Real-time dynamic monitoring: During the entire process, when certain thermal runaway categories or status conditions are met, the relevant steps will be repeated for continuous monitoring until the critical situation is accurately determined, ensuring the real-time status monitoring of the lithium battery pack and being able to capture thermal runaway-related changes in a timely manner.

[0079] (3) Improve response speed: By combining multiple parameters and characteristics for comprehensive judgment, compared with the traditional method of judging thermal runaway by relying on a single parameter (such as voltage or temperature alone), it can detect the trend of thermal runaway in lithium battery packs earlier and issue an early warning.

[0080] Embodiment 2 Based on the same inventive concept, Figure 2 As shown, the embodiment of the present invention further provides a lithium battery pack thermal runaway prediction device 200, comprising: The data acquisition module 210 is used to acquire the thermal imaging image, electrical parameters and cabin gas parameters of the lithium battery pack; An image recognition module 220 is used to perform image recognition on the thermal imaging image of the lithium battery pack to obtain features of the thermal imaging image of the lithium battery pack; The classification recognition module 230 is used to input the features of the thermal imaging image of the lithium battery pack into the lithium battery early warning classification model to obtain the thermal runaway category of the lithium battery pack; wherein the thermal runaway category includes: thermal runaway stage; A loop calling module 240, for returning to call the classification identification module 230 when the thermal runaway category of the lithium battery pack is the thermal runaway stage or when it is determined that the state of the lithium battery pack is the thermal runaway stage based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin; The target recognition module 250 is used to perform target detection and recognition on the thermal imaging image of the lithium battery pack using a target detection algorithm when the thermal runaway category of the lithium battery pack is the thermal runaway stage and based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, the state of the lithium battery pack is determined to be the thermal runaway stage, and determine the single cell with the highest temperature in the lithium battery pack.

[0081] Optionally, the image recognition module 220 is specifically used for: Determine the temperature value corresponding to each pixel of the thermal imaging image of the lithium battery pack; Determine pixels in the thermal imaging image of the lithium battery pack whose temperature value is greater than a preset temperature as marked pixels of each frame of the thermal imaging image, and determine the number of marked pixels; The total number of pixels of the thermal imaging image of the lithium battery pack is determined, and the ratio of the number of marked pixels to the total number of pixels is calculated to obtain the hot spot area ratio of the thermal imaging image of the lithium battery pack.

[0082] It should be understood that the device corresponds to the above-mentioned lithium battery pack thermal runaway prediction method embodiment and can execute the various steps involved in the above-mentioned method embodiment. The specific functions of the device can be found in the above description. To avoid repetition, the detailed description is appropriately omitted here. The device includes at least one software function module that can be stored in a memory in the form of software or firmware or solidified in the operating system (OS) of the device.

[0083] Embodiment 3 Based on the same inventive concept, an embodiment of the present invention further provides an electronic device, including: a processor and a memory, the memory storing machine-readable instructions executable by the processor, and the machine-readable instructions executing the above-mentioned lithium battery pack thermal runaway prediction method when executed by the processor.

[0084] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0085] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0086] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0087] Embodiment 4 Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium storing computer instructions. When the computer instructions are executed on a computer, the computer executes the above-mentioned method for predicting thermal runaway of a lithium battery pack.

[0088] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0089] The present application 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 application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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.

[0090] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] 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.

[0092] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0093] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0094] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0095] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for predicting thermal runaway of a lithium battery pack, characterized in that: include: S100, obtaining thermal imaging images, electrical parameters and cabin gas parameters of the lithium battery pack; S200, performing image recognition on the thermal imaging image of the lithium battery pack to obtain features of the thermal imaging image of the lithium battery pack; S300, inputting the features of the thermal imaging image of the lithium battery pack into a lithium battery early warning classification model to obtain a thermal runaway category of the lithium battery pack; wherein the thermal runaway category includes: a thermal runaway stage; S400, if the thermal runaway category of the lithium battery pack is the thermal runaway stage or based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, it is determined that the state of the lithium battery pack is the thermal runaway stage, then return to step S300; S500, if the thermal runaway category of the lithium battery pack is the thermal runaway stage and based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, it is determined that the state of the lithium battery pack is the thermal runaway stage, then a target detection algorithm is used to perform target detection and recognition on the thermal imaging image of the lithium battery pack to determine the single cell with the highest temperature in the lithium battery pack.

2. The method for predicting thermal runaway of a lithium battery pack according to claim 1, characterized in that: Thermal imaging images of lithium battery packs are obtained by: Get the temperature of the lithium battery pack; If the temperature of the lithium battery pack is lower than the first preset temperature, thermal imaging images of the lithium battery pack are intercepted at intervals of a first preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and one frame of thermal imaging images is randomly selected from the multiple frames of thermal imaging images of the lithium battery pack; If the temperature of the lithium battery pack is greater than the first preset temperature and less than the second preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the second preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting one frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; If the temperature of the lithium battery pack is greater than the second preset temperature and less than the third preset temperature, intercepting the thermal imaging image of the lithium battery pack at intervals of the third preset time length to obtain multiple frames of thermal imaging images of the lithium battery pack, and arbitrarily selecting one frame of thermal imaging image from the multiple frames of thermal imaging images of the lithium battery pack; Among them, the first preset temperature is less than the second preset temperature and less than the third preset temperature; the first preset time is less than the second preset time and less than the third preset time.

3. The method for predicting thermal runaway of a lithium battery pack according to claim 1, characterized in that: Before the step of performing image recognition on the thermal imaging image of the lithium battery pack to obtain the features of the thermal imaging image of the lithium battery pack, the method further includes: The thermal imaging image of the lithium battery pack is preprocessed; wherein the preprocessing includes: normalization processing, image enhancement processing, Gaussian filtering processing and histogram equalization processing.

4. The method for predicting thermal runaway of a lithium battery pack according to claim 1, characterized in that: Image recognition is performed on multiple frames of thermal imaging images of the lithium battery pack to obtain the features of the thermal imaging images of the lithium battery pack, including: Determine the temperature value corresponding to each pixel of the thermal imaging image of the lithium battery pack; Determine pixels in the thermal imaging image of the lithium battery pack whose temperature value is greater than a preset temperature as marked pixels of each frame of the thermal imaging image, and determine the number of marked pixels; The total number of pixels of the thermal imaging image of the lithium battery pack is determined, and the ratio of the number of marked pixels to the total number of pixels is calculated to obtain the hot spot area ratio of the thermal imaging image of the lithium battery pack.

5. The method for predicting thermal runaway of a lithium battery pack according to claim 1, characterized in that: The lithium battery warning classification model is trained in the following way: Obtain the hot spot area ratio of thermal imaging images of lithium battery packs of different thermal runaway categories; thermal runaway categories include: self-heating stage, thermal runaway stage and thermal runaway termination stage; The pre-built convolutional neural network model is trained using the hot spot area ratio of thermal imaging images of lithium battery packs of different thermal runaway categories. After the training is completed, a lithium battery early warning classification model is obtained.

6. The method for predicting thermal runaway of a lithium battery pack according to claim 1, characterized in that: The electrical parameters of the lithium battery pack include: voltage and current; the gas parameters in the cabin of the lithium battery pack include: the concentration of hydrogen in the cabin and the concentration of carbon monoxide in the cabin; Based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, determine the status of the lithium battery pack, including: If any one of the voltage, current, hydrogen concentration in the cabin, and carbon monoxide concentration of the lithium battery pack is greater than the corresponding preset value, it is determined that the state of the lithium battery pack is in the thermal runaway stage; If the voltage, current, cabin hydrogen concentration, and cabin carbon monoxide concentration of the lithium battery pack are all less than corresponding preset values, it is determined that the lithium battery pack is in a safe stage.

7. A device for predicting thermal runaway of a lithium battery pack, characterized in that: include: A data acquisition module, used to obtain thermal imaging images, electrical parameters and cabin gas parameters of the lithium battery pack; An image recognition module is used to perform image recognition on the thermal imaging image of the lithium battery pack to obtain the characteristics of the thermal imaging image of the lithium battery pack; The classification and recognition module is used to input the features of the thermal imaging image of the lithium battery pack into the lithium battery early warning classification model to obtain the thermal runaway category of the lithium battery pack; wherein the thermal runaway category includes: thermal runaway stage; A loop calling module, used for returning to call the classification identification module when the thermal runaway category of the lithium battery pack is the thermal runaway stage or when it is determined that the state of the lithium battery pack is the thermal runaway stage based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin; The target recognition module is used to perform target detection and recognition on the thermal imaging image of the lithium battery pack using a target detection algorithm when the thermal runaway category of the lithium battery pack is the thermal runaway stage and based on the electrical parameters of the lithium battery pack and the gas parameters in the cabin, the state of the lithium battery pack is determined to be the thermal runaway stage, and determine the single cell with the highest temperature in the lithium battery pack.

8. The device for predicting thermal runaway of a lithium battery pack according to claim 7, characterized in that: The image recognition module is specifically used for: Determine the temperature value corresponding to each pixel of the thermal imaging image of the lithium battery pack; Determine pixels in the thermal imaging image of the lithium battery pack whose temperature value is greater than a preset temperature as marked pixels of each frame of the thermal imaging image, and determine the number of marked pixels; The total number of pixels of the thermal imaging image of the lithium battery pack is determined, and the ratio of the number of marked pixels to the total number of pixels is calculated to obtain the hot spot area ratio of the thermal imaging image of the lithium battery pack.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores machine-readable instructions executable by the processor, and when the machine-readable instructions are executed by the processor, the method for predicting thermal runaway of a lithium battery pack according to any one of claims 1 to 7 is executed.

10. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on a computer, the computer is enabled to execute the method for predicting thermal runaway of a lithium battery pack according to any one of claims 1 to 7.

Images