Battery pack thermal runaway early warning method and system based on sensibilization optical fiber sensor

Abstract

The invention provides a battery pack thermal runaway early warning method and system based on a sensibilization optical fiber sensor, and belongs to the field of lithium battery thermal runaway early warning. According to the invention, the first fiber bragg grating used for measuring strain and the second fiber bragg grating used for measuring temperature are integrated, and the sensitivity coefficient is enabled to be less than 1 through size and mechanical parameter design so as to realize the sensibilization of the sensibilization fiber sensor, and the sensibilization fiber sensor is reasonably arranged. The surface temperature and strain data of the square lithium battery pack can be synchronously acquired without depending on a plurality of sensors, the problems of complex system, high cost and neglect of strain parameters caused by a plurality of sensors in the prior art are solved, and the high-sensitivity monitoring requirement of the weak strain at the initial stage of thermal runaway of the square lithium battery can be met; the problem that an existing sensor is insufficient in sensitivity is solved, data are processed in combination with a trained time sequence prediction model, the accuracy and reliability of thermal runaway early warning are finally improved, and thermal runaway of the lithium battery is effectively prevented.

Description

Technical Field

[0001] This invention belongs to the field of lithium battery thermal runaway early warning technology, specifically relating to a battery pack thermal runaway early warning method and system based on an enhanced fiber optic sensor. Background Technology

[0002] In today's society, with the rapid development of technology, lithium batteries are widely used in many fields, such as electric vehicles and portable electronic devices, due to their advantages such as high energy density and long cycle life. However, the safety of lithium batteries has always been a key factor restricting their further development and large-scale application. Battery safety has become a top priority in the field of lithium battery research.

[0003] Lithium-ion battery thermal runaway early warning, as a core technology for battery safety management, has received widespread attention. Traditional thermal runaway early warning schemes mostly employ external temperature threshold monitoring mechanisms, using temperature sensors deployed on the battery surface to collect real-time surface temperature data. When the detected temperature exceeds a pre-set threshold, the system triggers an alarm. This approach can provide early warning of lithium-ion battery thermal runaway to a certain extent and has the advantages of simple implementation and relatively controllable cost. In recent years, with the deepening of research on lithium-ion battery thermal runaway, researchers have begun to explore more advanced thermal runaway early warning methods, such as using multi-sensor fusion technology combined with multi-parameter data analysis, to identify early signs of thermal runaway earlier, while also adopting a tiered early warning strategy to improve the early warning effect.

[0004] Current detection methods often rely on multiple sensors to monitor multiple parameters, which complicates the system structure and increases costs. Furthermore, existing thermal runaway fault diagnosis models have shortcomings in parameter selection, often neglecting the crucial parameter of strain. For prismatic hard-shell lithium batteries, the strain generated on the surface in the early stages of thermal runaway is relatively weak, placing higher demands on sensor sensitivity, which existing sensors struggle to meet. In addition, the widely used transformer architecture suffers from computational bottlenecks in its global attention mechanism when processing thermal runaway warning data; as the number of variables increases, the inference speed slows down significantly, affecting the timeliness of warnings. Summary of the Invention

[0005] In view of this, the present invention provides a method and system for early warning of thermal runaway in battery packs based on an enhanced fiber optic sensor, aiming to improve the accuracy and reliability of early warning of thermal runaway and effectively prevent the occurrence of thermal runaway in lithium batteries.

[0006] To achieve the above objectives, the technical solution provided by the present invention is as follows:

[0007] In a first aspect, the present invention provides a method for early warning of thermal runaway of a battery pack based on a sensitive fiber optic sensor. The sensitive fiber optic sensor includes a first fiber grating for measuring strain and a second fiber grating for measuring temperature. The sensitivity coefficient of the sensitive fiber optic sensor is determined by the size and mechanical parameters of the sensor, and the size and mechanical parameters satisfy the condition that the sensitivity coefficient is less than 1, so as to achieve sensitivity enhancement of the measurement.

[0008] The method includes the following steps:

[0009] For square lithium battery packs, the battery side and electrode positions are selected as temperature strain monitoring points, and enhanced fiber optic sensors are used to cover the temperature strain monitoring points.

[0010] The enhanced fiber optic sensor is initialized and configured, and the surface temperature and strain data of the square lithium battery pack are collected through the first fiber grating and the second fiber grating of the enhanced fiber optic sensor.

[0011] Surface temperature and strain data are input into a trained time series prediction model to obtain temperature and strain prediction results, and thermal runaway early warning for square lithium battery packs is based on the prediction results.

[0012] Furthermore, the enhanced fiber optic sensor also includes: a circular elastic substrate;

[0013] The annular elastic substrate includes three sequentially connected annular rings, and a fixed end plate connected to the outer side of the two annular rings; the fixed end plate and the annular rings are horizontally distributed in sequence.

[0014] The lengths of the rings and the fixed end plate, as well as the connecting sections between the rings, are all the same, and the dimensions of the rings on both sides are identical.

[0015] The two ends of the first fiber grating are suspended and fixed on the middle ring along the horizontal direction;

[0016] The second fiber grating is fixed horizontally on the fixed end plate on the left side;

[0017] The first fiber grating and the second fiber grating have a preset wavelength interval.

[0018] Furthermore, the sensitivity coefficient of the enhanced fiber optic sensor is defined as the ratio of the strain of the sensor substrate to the axial strain of the first fiber grating.

[0019] Furthermore, the calculation expressions for the strain of the sensor substrate and the axial strain of the fiber grating are as follows:

[0020]

[0021]

[0022] In the formula, Strain on the substrate The axial strain of the first fiber grating; and These represent the horizontal length of the central ring and the change in length, respectively. and These represent the horizontal length and the change in length of the two circular rings, respectively. and These are the length of the connecting segment and the amount of length change, respectively. and These represent the horizontal length and length variation of the fixed-end plate, respectively.

[0023] Furthermore, the expression for calculating the sensitivity coefficient is as follows:

[0024]

[0025] In the formula, This is the sensitivity coefficient; and These are the outer and inner diameters of the central ring, respectively. and These represent the outer and inner diameters of the two circular rings, respectively; E is the elastic stiffness of the material; and d is the sensor thickness. For the stiffness of the connecting section; The stiffness of the fixed-end plate.

[0026] Furthermore, the calculation expressions for the surface temperature and strain data of the square lithium battery pack are as follows:

[0027]

[0028]

[0029] In the formula, and These represent the wavelength changes of the first fiber grating and the second fiber grating, respectively. and These are the temperature sensitivity coefficients of the first fiber grating and the second fiber grating, respectively. The strain sensitivity coefficient of the first fiber grating; and These represent the change in surface temperature and the change in strain of the first fiber grating, respectively.

[0030] Furthermore, enhanced fiber optic sensors are used to cover the temperature strain monitoring points, including:

[0031] Based on the distribution of temperature strain monitoring points, the laying path with the fewest fiber bending points is selected.

[0032] Based on the number of temperature strain monitoring points and the distance between each monitoring point, determine the total number of grating points that need to be written on the enhanced fiber optic sensor, as well as the spacing between adjacent grating points;

[0033] The engraved fiber optic sensor is fixed to the surface of the square lithium battery pack along the paving path.

[0034] Furthermore, the surface temperature and strain data are input into the trained time series prediction model to obtain the predicted temperature and strain results, including:

[0035] The surface temperature and strain data are input into the first prediction model to obtain the first prediction results of temperature and strain. The first prediction model is a convolutional neural network-bidirectional long short-term memory model pre-trained using the surface temperature and strain data of square lithium battery packs under normal operating conditions and thermal runaway fault conditions.

[0036] The first prediction result is used as the input of the second prediction model to obtain the second prediction result of temperature and strain; the second prediction model is a Mamba model pre-trained using historical data sequences of temperature and strain under thermal runaway fault conditions.

[0037] The second prediction result is used as the final prediction result to provide early warning of thermal runaway in square lithium battery packs.

[0038] Furthermore, the standard operating conditions are constant current charge-discharge cycle tests at 0.5C, 1C, and 1.5C rates conducted at room temperature;

[0039] Thermal runaway fault conditions include thermal runaway caused by internal short circuit, thermal runaway caused by overcharging, and thermal runaway caused by high temperature environment.

[0040] Secondly, the present invention provides a battery pack thermal runaway early warning system based on an enhanced fiber optic sensor, comprising:

[0041] The enhanced fiber optic sensor includes a first fiber grating for measuring strain and a second fiber grating for measuring temperature. The sensitivity coefficient of the enhanced fiber optic sensor is determined by the sensor's size and mechanical parameters, and the size and mechanical parameters satisfy the condition that the sensitivity coefficient is less than 1, so as to enhance the sensitivity of the measurement. For square lithium battery packs, the battery side and electrode positions are selected as temperature strain monitoring points, and the enhanced fiber optic sensor is used to cover the temperature strain monitoring points.

[0042] The data acquisition module is configured to initialize the sensitized fiber optic sensor and acquire surface temperature and strain data of the square lithium battery pack through the first and second fiber optic gratings of the sensitized fiber optic sensor.

[0043] The thermal runaway early warning module is used to input surface temperature and strain data into a trained time series prediction model to obtain temperature and strain prediction results, and to provide thermal runaway early warning for square lithium battery packs based on the prediction results.

[0044] In summary, this invention provides a battery pack thermal runaway early warning method and system based on an enhanced fiber optic sensor. By employing an enhanced fiber optic sensor that integrates a first fiber grating for strain measurement and a second fiber grating for temperature measurement, and whose sensitivity coefficient is designed to be less than 1 through size and mechanical parameter optimization, and by rationally deploying this enhanced fiber optic sensor, the surface temperature and strain data of a square lithium battery pack can be simultaneously acquired without relying on multiple sensors. This solves the problems of system complexity, high cost, and neglect of strain parameters caused by multiple sensors in existing technologies. Furthermore, it meets the high-sensitivity monitoring requirements for the weak strain in the early stages of thermal runaway in square lithium batteries, addressing the insufficient sensitivity of existing sensors. Combined with a trained time-series prediction model to process the data, the accuracy and reliability of thermal runaway early warning are ultimately improved, effectively preventing lithium battery thermal runaway. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 A flowchart of a battery pack thermal runaway early warning method based on an enhanced fiber optic sensor provided in an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram of the structure of the enhanced fiber optic sensor provided in an embodiment of the present invention;

[0048] Figure 3 This is a schematic diagram of the fiber optic sensor arrangement provided in an embodiment of the present invention.

[0049] In the attached diagram: 1-Second fiber Bragg grating, 2-Left fixed end plate, 3-Adhesive point, 4-First fiber Bragg grating, 5-Right fixed end plate, 6-Fiber optic cable, 7-Battery cell, 8-Positive electrode, 9-Safety valve, 10-Negative electrode, 11-Fiber Bragg grating sensor. Detailed Implementation

[0050] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0051] The technical terms of some of the prior art involved in this invention will be introduced below.

[0052] (1) Convolutional Neural Network (CNN): It is a feedforward neural network that automatically extracts local features of data through structures such as convolutional layers and pooling layers. It is widely used in image processing (such as image classification and object detection), speech recognition and other fields. Its core is to reduce the number of parameters by utilizing the weight sharing and local connectivity characteristics of convolutional kernels, while effectively capturing the spatial correlation of data.

[0053] (2) Bidirectional Long Short-Term Memory (BiLSTM): This is an improved version of LSTM (Long Short-Term Memory), consisting of a forward LSTM and a backward LSTM. It can simultaneously utilize historical and future information of a sequence and performs well in natural language processing (such as text sentiment analysis and machine translation) and time series prediction (such as stock price prediction and equipment failure prediction). It solves the gradient vanishing problem of traditional recurrent neural networks through gating mechanisms (input gate, forget gate, and output gate), and achieves effective capture of long sequence dependencies.

[0054] (3) Mamba model: This is a novel sequence model based on SSM (State Space Model), designed to address the computational complexity issue of the Transformer model when processing long sequences. The Mamba model achieves selective processing of input information and hardware-aware algorithm design by introducing a selection mechanism and a parallel scanning algorithm. This allows Mamba to maintain high efficiency during both training and inference, while achieving state-of-the-art performance in language and audio sequence modalities.

[0055] The various embodiments of the present invention will be described in detail below.

[0056] This embodiment provides a battery pack thermal runaway early warning method based on an enhanced fiber optic sensor. The enhanced fiber optic sensor includes a first fiber grating for measuring strain and a second fiber grating for measuring temperature. The sensitivity coefficient of the enhanced fiber optic sensor is determined by the sensor's size and mechanical parameters, and the size and mechanical parameters satisfy the condition that the sensitivity coefficient is less than 1, so as to enhance the sensitivity of the measurement.

[0057] It should be noted that the enhanced fiber optic sensor is a fiber optic sensing device that improves measurement sensitivity through structural design. It comprises two fiber gratings with different functions, capable of simultaneously monitoring both temperature and strain. The first fiber grating is a component specifically designed for strain measurement; its core principle is that the grating wavelength shifts with the deformation (tension / compression) of the object, and the strain value can be calculated by detecting the wavelength change. The second fiber grating is a component specifically designed for temperature measurement; its core principle is that the grating wavelength shifts with changes in ambient temperature, and the temperature value can be calculated by detecting the wavelength change. The sensitivity coefficient measures the sensor's responsiveness to changes in the measured parameter (temperature / strain). In this embodiment, the sensor's dimensions (e.g., substrate length) and mechanical parameters (e.g., material stiffness) are designed to make the coefficient less than 1, thereby amplifying minute changes in the measured parameter and achieving the enhanced sensitivity effect.

[0058] The method includes the following steps:

[0059] S11: For square lithium battery packs, the battery side and electrode positions are selected as temperature strain monitoring points, and enhanced fiber optic sensors are used to cover the temperature strain monitoring points.

[0060] It should be noted that when a square lithium battery experiences thermal runaway, the sides are prone to deformation (abnormal strain) due to internal gas expansion, while the electrodes become areas of abnormal temperature due to concentrated current and internal resistance heating. Selecting these two locations can more accurately capture early signals before thermal runaway and avoid monitoring blind spots.

[0061] The enhanced fiber optic sensor is tightly attached to the surface of the monitoring point to ensure that there is no gap between the sensor and the battery surface, reduce environmental interference (such as air insulation and vibration), and ensure the accuracy of temperature and strain data acquisition.

[0062] S12: Initialize and set the enhanced fiber optic sensor, and collect surface temperature and strain data of the square lithium battery pack through the first fiber grating and the second fiber grating of the enhanced fiber optic sensor.

[0063] It should be noted that by calibrating, interference from the initial state of the sensor (such as minor stress during installation and ambient reference temperature) is eliminated, and a zero-error measurement benchmark is established.

[0064] The first fiber grating detects the deformation of the battery surface and converts the strain change into a wavelength shift signal; the second fiber grating detects the battery surface temperature and converts the temperature change into a wavelength shift signal; the two signals are transmitted synchronously and converted into specific temperature and strain values.

[0065] S13: Input the surface temperature and strain data into the trained time series prediction model to obtain the predicted results of temperature and strain, and perform thermal runaway early warning for square lithium battery packs based on the prediction results.

[0066] It should be noted that time series forecasting models are algorithmic models that learn the patterns of change based on historical time series data (such as continuously collected temperature / strain data) and then predict the trend of parameter changes over a future period of time.

[0067] The collected continuous temperature and strain data (time-series data) are input into the trained model. The model learns from the normal and abnormal patterns of change in historical data (such as the precursor characteristics of sudden temperature rises and strain increases) and outputs the parameter prediction results for a future period of time. A preset thermal runaway risk threshold is set (such as the predicted temperature exceeding 80°C or the predicted strain exceeding the material limit). When the predicted results output by the model reach or approach the threshold, the system automatically triggers an early warning, prompting staff to intervene.

[0068] This embodiment provides a battery pack thermal runaway early warning method based on an enhanced fiber optic sensor. This method overcomes the problem of insufficient sensitivity in traditional fiber optic sensors by setting the sensor's size and mechanical parameters to achieve a sensitivity coefficient less than 1, enabling it to capture minute temperature / strain changes before battery thermal runaway. Secondly, considering the structural characteristics of square batteries, the method precisely selects the sides (strain-sensitive areas) and electrodes (temperature-sensitive areas) as monitoring points, avoiding signal lag or ineffective monitoring caused by traditional random point placement. Finally, it abandons the passive mode of alarming only when real-time data exceeds limits, and instead uses a time series prediction model to predict parameter change trends in advance, significantly extending the early warning window for thermal runaway and gaining crucial time for battery safety intervention.

[0069] Please see Figure 1 , Figure 1 This is an implementation flow of a battery pack thermal runaway early warning method based on an enhanced fiber optic sensor, designed based on the above embodiments. The flow is divided into two parts: model training and actual prediction. In the model training phase, experiments are conducted and data is collected under normal and fault conditions. The collected temperature and strain data are input into a CNN-BiLSTM model, and then its output data is input into a Mamba model to complete preliminary training. In the actual prediction phase, fiber optic sensors are first installed on the surface of the battery pack and their parameters are initialized. Then, the surface temperature and strain data of the battery are collected in real time and input into the trained Mamba model for prediction. The following describes the process in conjunction with... Figure 1Some other embodiments of the present invention will be further described.

[0070] like Figure 2 As shown, in one embodiment of the present invention, the enhanced fiber optic sensor further includes: a circular elastic substrate;

[0071] The annular elastic substrate includes three rings connected in sequence, and fixed end plates (such as left fixed end plate 2 and right fixed end plate 5, both with a horizontal length of L4) connected to the outer sides of the two rings; the fixed end plates and the rings are horizontally distributed in sequence.

[0072] The lengths of the ring and the fixed end plate, as well as the connecting sections between the rings, are all the same (all L3), and the dimensions of the two rings are consistent (the horizontal length of both rings is L2, and the horizontal length of the middle ring is L1).

[0073] The two ends of the first fiber grating 4 are suspended and fixed on the middle ring along the horizontal direction;

[0074] The second fiber grating 1 is fixed horizontally on the fixed end plate on the left side;

[0075] The first fiber grating 4 and the second fiber grating 1 have a preset wavelength interval.

[0076] The sensor mainly consists of two FBGs (Fiber Bragg Gratings) and a circular elastic substrate. The two ends of FBG1 are fixed to the large central ring with adhesive, while FBG2 is directly attached to the fixed end plate on the left.

[0077] Because FBG1 is suspended and attached to the large central ring, strain transfer loss that might occur with direct attachment is avoided. Therefore, the strain it measures is the actual strain between the two attachment points 3. When both ends of the sensor are fixed to the object being measured, the deformation of the object is transmitted to the ring structure through the fixed ends, causing deformation of the central ring and consequently a change in the wavelength of FBG1. Conversely, the left side plate does not deform and is therefore unaffected by strain. This means that FBG2 attached here is only affected by temperature changes and can be used for temperature compensation. To effectively distinguish fiber optic gratings performing different functions, this embodiment uses two FBGs with a certain wavelength interval during the packaging process.

[0078] In a further embodiment of the present invention, the sensitivity coefficient of the enhanced fiber optic sensor is defined as the ratio of the strain of the sensor substrate to the axial strain of the fiber optic grating.

[0079] by Figure 2Taking the enhanced fiber optic sensor shown as an example, the sensor's sensitivity is mainly determined by the ratio between the length of the ring structure and the sensor itself. Assuming the strain transfer between the fiber grating and the elastic substrate is a rigid connection, when the sensor structure is subjected to an external force F, the large central ring structure possesses stiffness. and length The change in its length after being subjected to force is The stiffness of the two side structures is , length is The change in length after being subjected to force is The stiffness at the connection section is... , length is The change in length after being subjected to force is The stiffness of the two-panel structure is The change in length after being subjected to force is From the mechanics of materials, we can obtain:

[0080]

[0081] Let be the axial strain of the FBG. Assuming the strain is the entire substrate, in a further embodiment of the present invention, the calculation expressions for the strain of the sensor substrate and the axial strain of the fiber optic grating are respectively:

[0082]

[0083]

[0084] In the formula, Strain on the substrate The axial strain of the fiber grating; and These represent the horizontal length of the central ring and the change in length, respectively. and These represent the horizontal length and the change in length of the two circular rings, respectively. and These are the length of the connecting segment and the amount of length change, respectively. and These represent the horizontal length and length variation of the fixed-end plate, respectively.

[0085] Therefore, the sensitivity coefficient K is:

[0086]

[0087] When K < 1, the sensitivity enhancement of fiber optic strain measurement can be achieved; the smaller K is, the better the sensitivity enhancement effect.

[0088] According to mechanics of materials, the stiffness of the ring structure and the stiffness of the two side structures can be expressed as:

[0089]

[0090]

[0091]

[0092]

[0093] Where E is the elastic stiffness of the material. and These are the outer and inner diameters of the central ring, respectively. and Let d be the outer diameter and inner diameter of the two circular rings, respectively; d be the thickness of the sensor; and A be the cross-sectional area.

[0094] In a further embodiment of the present invention, the sensitivity coefficient K can be ultimately simplified to:

[0095]

[0096] In the formula, This is the sensitivity coefficient; and These are the outer and inner diameters of the central ring, respectively. and These represent the outer and inner diameters of the two circular rings, respectively; E is the elastic stiffness of the material; and d is the sensor thickness. For the stiffness of the connecting section; The stiffness of the fixed-end plate.

[0097] In one embodiment of the present invention, the calculation expressions for the surface temperature and strain data of the square lithium battery pack are as follows:

[0098]

[0099]

[0100] In the formula, and These are the wavelength changes of the first fiber grating 4 and the second fiber grating 1, respectively. and These are the temperature sensitivity coefficients of the first fiber grating 4 and the second fiber grating 1, respectively. The strain sensitivity coefficient of the first fiber grating 4; and These represent the change in surface temperature and the change in strain of the first fiber grating 4, respectively.

[0101] In one embodiment of the present invention, a sensitive fiber optic sensor is used to cover the temperature strain monitoring point, including:

[0102] S21: Based on the distribution of temperature strain monitoring points, select the laying path with the fewest fiber bending points;

[0103] S32: Based on the number of temperature strain monitoring points and the distance between each monitoring point, determine the total number of grating points that need to be written on the sensitive fiber optic sensor, as well as the spacing between adjacent grating points;

[0104] S33: Fix the engraved sensitive fiber optic sensor onto the surface of the square lithium battery pack along the laying path.

[0105] Please see Figure 3 , Figure 3 An arrangement of fiber optic sensors is shown. Among them, Figure 3 The left figure shows the arrangement of the grating sensor 11 along the direction of the arrangement of multiple battery cells 7, so that the grating sensor 11 on the optical fiber 6 is arranged in sequence close to the electrode position (the electrode position is determined according to the position of the positive electrode 8, negative electrode 10 and safety valve 9); the right figure shows the arrangement of the grating sensor 11 on the optical fiber 6 along the direction of the arrangement of multiple battery cells 7, so that the grating sensor 11 is arranged in sequence close to the side position of the battery.

[0106] In one embodiment of the present invention, surface temperature and strain data are input into a trained time series prediction model to obtain prediction results for temperature and strain, including:

[0107] S41: Input the surface temperature and strain data into the first prediction model to obtain the first prediction results of temperature and strain; the first prediction model is a convolutional neural network-bidirectional long short-term memory model pre-trained using the surface temperature and strain data of the square lithium battery pack under normal operating conditions and thermal runaway fault conditions.

[0108] For battery packs with integrated sensors, routine operating condition tests and thermal runaway fault condition tests are conducted, including thermal runaway caused by internal short circuits, thermal runaway caused by overcharging, and thermal runaway caused by high temperature. The temperature and strain distribution of the battery pack are collected during the test.

[0109] The collected temperature and strain data are used as the input sequence of the CNN. Multiple features are extracted through multiple convolutional layers in the CNN. Each convolutional layer sets a convolutional kernel of different size for different feature mappings of the input sequence. The temperature and strain sequences output by the CNN are used as the input of the BiLSTM.

[0110] The temperature data update process of BiLSTM is represented as follows:

[0111]

[0112] In the formula, This indicates the positive output of the LSTM in time. This represents the backward output of the LSTM in time. Indicates the predicted temperature. This represents the output of the BiLSTM. The updating of strain data is similar.

[0113] S42: Use the first prediction result as the input of the second prediction model to obtain the second prediction result of temperature and strain; the second prediction model is a Mamba model pre-trained using historical data sequences of temperature and strain under thermal runaway fault conditions.

[0114] The predicted temperature and strain obtained from the CNN-BiLSTM model are used as the input sequence of the Manba model, and the processed... , The input data is divided into (S, L, D) parts, where S represents the batch size, L represents the sequence length, and D represents the input dimension. A selective state-space model is constructed, which includes (Δ, A, B, C) parameters, a discretized linear recursive equation, and a convolution calculation equation. The discretized linear recursive equation is expressed as:

[0115]

[0116] In the formula, Δ represents the step size, A represents the system matrix, B represents the control matrix, and I represents the identity matrix. Represents the discretized system matrix. Represents the discretized control matrix;

[0117] The convolution calculation equation is expressed as:

[0118]

[0119]

[0120] In the formula, Let C represent the hidden state at time t, and let C represent the output matrix. This represents the output of the Mabam model.

[0121] S43: Use the second prediction result as the final prediction result to provide early warning of thermal runaway in square lithium battery packs.

[0122] In a further embodiment of the present invention, the conventional operating condition is constant current charge-discharge cycle testing at 0.5C, 1C and 1.5C rates conducted at room temperature;

[0123] Thermal runaway fault conditions include thermal runaway caused by internal short circuit, thermal runaway caused by overcharging, and thermal runaway caused by high temperature environment.

[0124] Based on the same inventive concept, this application also provides a battery pack thermal runaway early warning device based on an enhanced fiber optic sensor for implementing the aforementioned battery pack thermal runaway early warning method based on an enhanced fiber optic sensor. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in the embodiments of the battery pack thermal runaway early warning device based on an enhanced fiber optic sensor provided below can be found in the limitations of the battery pack thermal runaway early warning method based on an enhanced fiber optic sensor described above, and will not be repeated here.

[0125] This invention provides a battery pack thermal runaway early warning system based on an enhanced fiber optic sensor, comprising:

[0126] The enhanced fiber optic sensor includes a first fiber grating for measuring strain and a second fiber grating for measuring temperature. The sensitivity coefficient of the enhanced fiber optic sensor is determined by the sensor's size and mechanical parameters, and the size and mechanical parameters satisfy the condition that the sensitivity coefficient is less than 1, so as to enhance the sensitivity of the measurement. For square lithium battery packs, the battery side and electrode positions are selected as temperature strain monitoring points, and the enhanced fiber optic sensor is used to cover the temperature strain monitoring points.

[0127] The data acquisition module is configured to initialize the sensitized fiber optic sensor and acquire surface temperature and strain data of the square lithium battery pack through the first and second fiber optic gratings of the sensitized fiber optic sensor.

[0128] The thermal runaway early warning module is used to input surface temperature and strain data into a trained time series prediction model to obtain temperature and strain prediction results, and to provide thermal runaway early warning for square lithium battery packs based on the prediction results.

[0129] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0130] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0131] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0132] In the embodiments disclosed in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0133] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for early warning of thermal runaway in battery packs based on an enhanced fiber optic sensor, characterized in that, The enhanced fiber optic sensor includes a first fiber grating for measuring strain and a second fiber grating for measuring temperature. The sensitivity coefficient of the enhanced fiber optic sensor is determined by the sensor's size and mechanical parameters, and the size and mechanical parameters satisfy a sensitivity coefficient less than 1 to achieve enhanced measurement sensitivity. The method includes the following steps: For square lithium battery packs, the battery side and electrode positions are selected as temperature strain monitoring points, and the enhanced fiber optic sensor is used to cover the temperature strain monitoring points. The enhanced fiber optic sensor is initialized and configured, and the surface temperature and strain data of the square lithium battery pack are collected through the first fiber grating and the second fiber grating of the enhanced fiber optic sensor. The surface temperature and strain data are input into a trained time series prediction model to obtain temperature and strain prediction results, and thermal runaway early warning for square lithium battery packs is performed based on the prediction results.

2. The method for early warning of thermal runaway of battery packs based on an enhanced fiber optic sensor according to claim 1, characterized in that, The enhanced fiber optic sensor further includes: a circular elastic substrate; The annular elastic substrate includes three sequentially connected annular rings and a fixed end plate connected to the outer sides of the two annular rings; the fixed end plate and the annular rings are horizontally distributed in sequence. The lengths of the ring, the fixed end plate, and the connecting sections between the rings are all the same, and the dimensions of the rings on both sides are identical. The two ends of the first fiber grating are suspended and fixed on the middle ring along the horizontal direction; The second fiber grating is fixed horizontally to the fixed end plate on the left side; The first fiber grating and the second fiber grating have a preset wavelength interval.

3. The method for early warning of thermal runaway of battery packs based on an enhanced fiber optic sensor according to claim 2, characterized in that, The sensitivity coefficient of the enhanced fiber optic sensor is defined as the ratio of the strain of the sensor substrate to the axial strain of the first fiber grating.

4. The battery pack thermal runaway early warning method based on an enhanced fiber optic sensor according to claim 3, characterized in that, The calculation expressions for the strain of the sensor substrate and the axial strain of the first fiber grating are as follows: In the formula, Strain on the substrate The axial strain of the first fiber grating; and These represent the horizontal length of the central ring and the change in length, respectively. and These represent the horizontal length and the change in length of the two circular rings, respectively. and These are the length of the connecting segment and the amount of length change, respectively. and These represent the horizontal length and length variation of the fixed-end plate, respectively.

5. The battery pack thermal runaway early warning method based on an enhanced fiber optic sensor according to claim 4, characterized in that, The expression for calculating the sensitivity coefficient is as follows: In the formula, The sensitivity coefficient is mentioned above; and These are the outer and inner diameters of the central ring, respectively. and These represent the outer and inner diameters of the two circular rings, respectively; E is the elastic stiffness of the material; and d is the sensor thickness. For the stiffness of the connecting section; The stiffness of the fixed-end plate.

6. The method for early warning of thermal runaway of a battery pack based on an enhanced fiber optic sensor according to claim 1, characterized in that, The calculation expressions for the surface temperature and strain data of the square lithium battery pack are as follows: In the formula, and These are the wavelength changes of the first fiber grating and the second fiber grating, respectively. and These are the temperature sensitivity coefficients of the first fiber grating and the second fiber grating, respectively; The strain sensitivity coefficient of the first fiber grating; and These represent the change in surface temperature and the change in strain of the first fiber grating, respectively.

7. The method for early warning of thermal runaway of a battery pack based on an enhanced fiber optic sensor according to claim 1, characterized in that, The enhanced fiber optic sensor is used to lay the fiber at the temperature strain monitoring point, including: Based on the distribution of the temperature strain monitoring points, the laying path with the fewest fiber bending points is selected. Based on the number of temperature strain monitoring points and the distance between each monitoring point, determine the total number of grating points to be engraved on the enhanced fiber optic sensor, as well as the spacing between adjacent grating points; The engraved fiber optic sensor is fixed to the surface of the square lithium battery pack along the laying path.

8. The method for early warning of thermal runaway of a battery pack based on an enhanced fiber optic sensor according to claim 1, characterized in that, The surface temperature and strain data are input into a trained time series prediction model to obtain the predicted temperature and strain results, including: The surface temperature and strain data are input into the first prediction model to obtain the first prediction results of temperature and strain; the first prediction model is a convolutional neural network-bidirectional long short-term memory model pre-trained using the surface temperature and strain data of the square lithium battery pack under normal operating conditions and thermal runaway fault conditions. The first prediction result is used as the input of the second prediction model to obtain the second prediction result of temperature and strain; the second prediction model is a Mamba model pre-trained using historical data sequences of temperature and strain under thermal runaway fault conditions. The second prediction result is used as the final prediction result to provide early warning of thermal runaway in square lithium battery packs.

9. The method for early warning of thermal runaway of a battery pack based on an enhanced fiber optic sensor according to claim 8, characterized in that, The standard operating conditions are constant current charge-discharge cycle tests at 0.5C, 1C, and 1.5C rates conducted at room temperature. The thermal runaway fault conditions include thermal runaway caused by internal short circuit, thermal runaway caused by overcharging, and thermal runaway caused by high temperature environment.

10. A battery pack thermal runaway early warning system based on an enhanced fiber optic sensor, characterized in that, include: The enhanced fiber optic sensor includes a first fiber grating for measuring strain and a second fiber grating for measuring temperature. The sensitivity coefficient of the enhanced fiber optic sensor is determined by the sensor's size and mechanical parameters, and the size and mechanical parameters satisfy the condition that the sensitivity coefficient is less than 1, thereby enhancing the sensitivity of the measurement. For a square lithium battery pack, the battery side and electrode positions are selected as temperature strain monitoring points, and the enhanced fiber optic sensor is used to cover these temperature strain monitoring points. The data acquisition module is configured to initialize the enhanced fiber optic sensor and acquire the surface temperature and strain data of the square lithium battery pack through the first fiber grating and the second fiber grating of the enhanced fiber optic sensor. The thermal runaway early warning module is used to input the surface temperature and strain data into a trained time series prediction model to obtain the predicted results of temperature and strain, and to provide thermal runaway early warning for the square lithium battery pack based on the predicted results.

Images