Early warning method and early warning system for thermal runaway of lithium ion battery

Abstract

This invention belongs to the field of lithium-ion battery safety monitoring and discloses an early warning method and system for lithium battery thermal runaway. This warning system and method possess high sensitivity and advantages such as in-situ monitoring, rapid response, and strong anti-interference. By directly capturing the molecular fingerprint information of the thermal decomposition of the electrolyte inside the battery, it achieves a leap from monitoring external physical parameters to identifying internal chemical mechanisms. This invention can accurately identify internal chemical anomaly signals during the initiation stage of battery thermal runaway, and has advantages such as fast response speed, no monitoring lag, and high sensitivity. Compared with traditional monitoring technologies that rely on external temperature or gas, it can provide early warning several minutes to tens of minutes in advance, providing a highly reliable real-time diagnostic solution for the safety management of power batteries and energy storage systems.

Description

Technical Field

[0001] This invention belongs to the field of lithium-ion battery safety monitoring, specifically relating to an early warning method and system for thermal runaway of lithium-ion batteries based on infrared evanescent wave spectroscopy analysis. Background Technology

[0002] With the advancement of global energy structure transformation, the new energy vehicle and large-scale energy storage industries have entered a period of rapid development. Lithium-ion batteries, with their superior safety and stability, long cycle life, and significant cost advantages, have become the mainstream choice for energy storage solutions in the current market, widely used in various end-user scenarios such as the automotive industry and grid-scale energy storage stations. However, under conditions of thermal abuse such as overcharging, short circuits, high-temperature environments, or mechanical shocks, lithium-ion batteries still face a severe risk of thermal runaway. This failure process exhibits a typical characteristic of "hidden in the early stages and explosive in the later stages": its initial stage (usually around 200°C) mainly involves microscopic chemical reactions such as early thermal decomposition of the electrolyte. During this stage, the external temperature change of the battery is extremely insignificant, and the heat generated by internal chemical instability is often masked by the battery's huge heat capacity, making it difficult to detect using conventional detection methods. This is precisely the critical time window for achieving accurate early warning of thermal runaway and preventing chain reactions. If timely identification and intervention are not implemented during this hidden stage, a violent exothermic chain reaction will rapidly occur inside the battery, potentially leading to serious safety accidents such as battery fire, combustion, or even explosion.

[0003] Existing lithium-ion battery thermal runaway monitoring technologies have evolved from external parameter monitoring to gas sensing, and then to electrochemical state assessment. However, these technologies all have fundamental limitations when applied to early warning of thermal runaway, mainly including: 1) Lag in physical parameter detection: Traditional temperature sensors (such as thermistors and thermocouples) and voltage monitoring units mainly rely on changes in external signals of the battery. Due to the significant thermal resistance and conduction delay between the battery casing and the internal active materials, when the external temperature shows an abnormal rise, the battery has usually already entered an irreversible and violent heat release phase, resulting in a significant lag in the warning signal.

[0004] 2) Gas detection response delay: Early warning methods based on gas sensors rely on the accumulation of internal battery pressure until the pressure relief valve opens, or on the diffusion of electrolyte decomposition byproducts onto the sensor surface. This process is affected by the battery packaging structure, venting path, and sensor installation location, resulting in a long response time that is difficult to meet millisecond-level safety protection requirements.

[0005] 3) Complexity of electrochemical and acoustic monitoring: Although electrochemical impedance spectroscopy (EIS) or acoustic ultrasonic monitoring technology is sensitive to changes in internal structure, it is easily affected by battery cycle aging, complex mechanical vibration environment and strong interference from electrochemical background noise in practical applications, resulting in limited sensitivity.

[0006] Therefore, there is an urgent need to develop an early warning method and system for lithium battery thermal runaway electrolyte to overcome the time delay bottleneck of traditional early warning technology and meet the extreme requirements of the new energy industry for battery safety. Summary of the Invention

[0007] To address the problems existing in the background technology, this invention provides an early warning method and system for lithium battery thermal runaway. This early warning system and method are highly sensitive and possess advantages such as in-situ monitoring, rapid response, and strong anti-interference capabilities. By directly capturing the molecular fingerprint information of the thermal decomposition of the electrolyte inside the battery, it achieves a leap from monitoring external physical parameters to identifying internal chemical mechanisms, providing a reliable solution for early screening and safety protection of lithium battery thermal runaway.

[0008] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a monitoring device for monitoring the inside of a battery, comprising an infrared light source, a ZnSe focusing lens, an infrared fiber evanescent wave sensor, and a detector arranged sequentially. A beam of infrared light emitted by the infrared light source is focused and coupled into the infrared fiber evanescent wave sensor through the ZnSe lens, generating an evanescent wave at the interface between the fiber and the battery's internal medium, and interacting with the electrolyte and its decomposition products coated near the waist region. The detector receives the optical signal from the fiber and converts it into an evanescent wave infrared absorption spectrum. The infrared fiber evanescent wave sensor includes a tapered chalcogenide micro / nano fiber optic sensing element that penetrates the battery casing and is embedded in situ inside the battery. The conical chalcogenide micro / nano optical sensing element is a bare-core structure, fabricated using Ge-As-Se-Te chalcogenide glass, and can transmit signals in the 800-4000 cm⁻¹ range. -1 Mid-infrared light, with infrared transmission band covering 2-16 μm; The conical chalcogenide micro / nano optical sensing element is fabricated from optical fibers with a diameter of 150-1000 μm, and includes a transition region and a waist region, with the waist region diameter controlled between 100-200 μm. The conical sulfide-based micro / nano optical sensing element is embedded in situ inside the battery model and hermetically sealed. This hermetically sealed design ensures no electrolyte leakage during charge-discharge cycles.

[0009] According to the above scheme, the Ge-As-Se-Te chalcogenide glass is preferably Ge. 20 As 20 Se 15 Te 45 .

[0010] According to the above scheme, the waist length of the optical fiber is 10-30 mm.

[0011] According to the above scheme, the detector is an MCT liquid nitrogen-cooled detector.

[0012] Secondly, this invention provides the application of the aforementioned monitoring device in early warning of battery thermal runaway. Based on the characteristic absorption peaks of the main electrolyte components, such as ethylene carbonate (EC) or dimethyl carbonate (DMC), early warning of battery thermal runaway is performed using solvent consumption kinetics.

[0013] Thirdly, the present invention provides an early warning method for battery thermal runaway based on solvent consumption kinetics, the method comprising: S1 Based on the above monitoring device, in-situ real-time infrared evanescent wave spectral data of electrolyte solvent components are collected; S2 combines the electrolyte solvent component concentration calibration curve to invert the infrared evanescent wave spectral signal into the electrolyte solvent component concentration, and constructs the concentration-temperature correlation curve of electrolyte solvent component evolution with battery internal temperature; S3 analyzes the concentration-temperature correlation curve of electrolyte solvent components as the battery internal temperature evolves, and combines the dynamic consumption rate of solvent components at different reaction stages to obtain the solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold. S4 When the system detects that the consumption rate of the battery solvent components exceeds the preset dynamic threshold, it determines that the internal chemical balance of the battery has been broken and enters the thermal runaway initiation stage.

[0014] According to the above scheme, based on the characteristic absorption peaks of the electrolyte solvent components in S2, and combined with the Lambert-Beer law, a linear quantitative relationship between absorbance and concentration is established to obtain the electrolyte solvent component concentration calibration curve.

[0015] According to the above scheme, the solvent dynamic consumption rate curve is obtained by taking the first derivative of the concentration-battery temperature correlation curve of the solvent component. The solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold, is set by analyzing the solvent dynamic consumption rate curve.

[0016] According to the above scheme, the warning threshold includes a primary dynamic threshold k1 and a secondary threshold k2; the primary dynamic threshold k1 is the statistical upper limit value of the solvent dynamic consumption rate curve in the steady-state fluctuation stage, which is used to characterize the critical limit of the monitoring signal leaving the normal background noise and entering the early abnormal deviation stage; the secondary threshold k2 is the kinetic step starting point value of the solvent dynamic consumption rate curve from the steady evolution to the rapid acceleration stage, which is used to characterize the burst characteristic point of irreversible and violent thermal degradation of the electrolyte.

[0017] According to the above scheme, the early warning threshold is obtained based on the dynamic consumption rate analysis of electrolyte components ethylene carbonate (EC) or dimethyl carbonate (DMC).

[0018] According to the above scheme, an early warning threshold is obtained based on the dynamic consumption rate analysis of the electrolyte component ethylene carbonate (EC). The early warning threshold includes a primary dynamic threshold and a secondary threshold, wherein the primary dynamic threshold k1 is 1.5-2 mmol / (L). ℃); the secondary threshold k2 is 3.5-4 mmol / (L) ℃).

[0019] According to the above scheme, characteristic stretching vibration peaks are selected as response signals based on molecular fingerprint characteristics. For the electrolyte solvent component dimethyl carbonate (DMC), the response signal is based on the C=O stretching vibration peak at 1754 cm⁻¹. -1 Quantitative and quantitative analysis of DMC was performed. For the electrolyte solvent component ethylene carbonate (EC), the characteristic absorption peak of ethylene carbonate (EC) at 1803 cm⁻¹ was determined. -1 Perform quantitative analysis of EC.

[0020] Fourthly, the present invention provides a lithium battery thermal runaway early warning system, comprising: The detection module is used to collect infrared evanescent wave spectral data of electrolyte solvent components in situ in real time based on an infrared fiber optic evanescent wave sensor. The analysis and processing module is used to combine the electrolyte solvent component concentration calibration curve, invert the infrared evanescent wave spectral signal into the electrolyte solvent component concentration, and construct the concentration-temperature correlation curve of the electrolyte solvent component as the battery internal temperature evolves. The warning threshold module is used to analyze and process the concentration-temperature correlation curve of the electrolyte solvent component as the internal temperature of the battery, and combine the dynamic consumption rate of the solvent component at different reaction stages to set the solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold. The determination module is used to determine that the internal chemical balance of the battery has been broken and the battery has entered the thermal runaway initiation stage when the consumption rate of the battery solvent component exceeds the preset dynamic threshold.

[0021] The aforementioned early warning systems or methods are applied in new energy vehicles, safe energy storage power stations, and battery health status assessment.

[0022] This invention quantifies the kinetics of electrolyte decomposition and interfacial reactions by real-time acquisition and analysis of the characteristic absorption peak evolution of the molecular fingerprint region, thereby constructing a quantitative early warning index for thermal runaway risk. This invention can accurately identify internal chemical anomalies during the battery thermal runaway initiation stage (≤200℃), offering advantages such as fast response, no monitoring lag, and high sensitivity. Compared to traditional monitoring technologies that rely on external temperature or gas, it can provide early warnings several minutes to tens of minutes in advance, offering a highly reliable real-time diagnostic solution for the safety management of power batteries and energy storage systems.

[0023] The beneficial effects of this invention are: 1) The infrared fiber evanescent wave sensor and monitoring device for battery internal monitoring provided by the present invention are used for electrolyte component detection. It has high sensitivity and signal-to-noise ratio: the detection sensitivity for DMC and EC can reach 0.012 au / (mmol / L) and 0.023 au / (mmol / L) respectively, and it has excellent corrosion resistance and long-term stability.

[0024] 2) The early warning method and system provided by this invention can reveal the microscopic mechanism of failure: by utilizing the characteristics of "molecular fingerprint", the concentration evolution trend of electrolyte solvent components (such as EC and DMC) can be quantified in real time, and the dynamic process of solvent component consumption and decomposition can be truly presented, providing high-precision in-situ data support for battery aging research and thermal runaway risk assessment.

[0025] 3) The infrared fiber optic evanescent wave sensor and monitoring device for internal battery monitoring provided by this invention: The chalcogenide fiber is highly flexible and has a small diameter, making it easy to embed in situ without altering the battery structure, and possesses... High sensitivity, high reliability sex, Intelligent safety monitoring system suitable for electric vehicles and large-scale energy storage power stations.

[0026] 4) The early warning method and system provided by the present invention can be used for early warning of thermal runaway of lithium batteries, and the early warning capability is significantly enhanced: when the battery surface temperature is only 40-60 ℃ and the traditional sensor has not yet alarmed, the system can identify the signs of thermal runaway by capturing the chemical signal of thermal decomposition of the internal electrolyte, and the warning time can be advanced by several minutes to tens of minutes. Attached Figure Description

[0027] Figure 1 This is a flowchart of the experimental preparation process in Embodiment 1 of the present invention. 1 is an optical fiber preform, 2 is a cross-sectional view of the optical fiber, 3 is a tapered optical fiber, and 4 is a diagram of the internal structure of the optical fiber embedded in the battery. Figure 2 This is a schematic diagram of the infrared fiber evanescent wave sensor structure in Embodiment 1 of the present invention. 32 is the transition region and 33 is the waist region.

[0028] Figure 3 This is a diagram of the infrared fiber optic evanescent wave sensing system in Embodiment 2 of the present invention; Figure 4 The evanescent infrared absorption spectra of the electrolyte inside the battery at different temperatures in Example 3 of this invention; Figure 5The infrared absorption spectra of solutions of different concentrations (10-500 mmol / L) detected by the fiber optic sensor in Example 4 are shown in Figure a, which is the dimethyl carbonate (DMC) solution and Figure D, which is the ethylene carbonate (EC) solution. Anhydrous acetonitrile was used as the solvent in both cases. Figure 6 Example 4 compares the detection sensitivity of the sensor to two different solutions; Figure 7 Comparison of evanescent infrared absorption spectra of the infrared fiber optic sensor in Example 5 after different usage times; Figure 8 This is a comparison of the sensing performance of the infrared fiber optic sensor in Embodiment 5 of the present invention after high-temperature cycling; Figure 9 The evanescent infrared absorption difference spectrum of the battery inside Example 6 at different temperatures is shown, with the spectrum at 30 °C as the background. Figure 10 This is a graph showing the relationship between temperature and concentration established in Example 6 of the present invention; Figure 11 This is a characteristic curve of the dynamic consumption rate of electrolyte solvent |dC / dT| as a function of temperature in Example 6 of the present invention. Figure 12 This is a comparison chart of the early warning system of this invention and traditional monitoring methods. Detailed Implementation

[0029] The principles and features of the present invention are described below with reference to the accompanying drawings and specific embodiments. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0030] Example 1: Fabrication and in-situ integration of a chalcogenide tapered micro / nano fiber optic sensor like Figure 1 As shown, this embodiment provides an infrared fiber optic evanescent wave sensor for monitoring the inside of a battery, including a battery model 4 (Swagelok battery model) with a reaction cavity and a tapered micro / nano fiber 3 that penetrates the battery model and is embedded in situ in the electrolyte. The portion of the tapered micro / nano fiber 3 located inside the battery is formed into a tapered structure by drawing a tapered shape. The tapered micro / nano fiber 3 is formed by heating and drawing the middle section of a chalcogenide infrared fiber 2, which is fabricated using a glass preform 1 under a fiber drawing tower. The structure of the tapered micro / nano fiber 3 is as follows. Figure 2 As shown, the structure includes a transition region 32 and a waist region 33 (i.e., a sensing region). The waist region of the tapered structure has a diameter of 100-200 μm and a length of 15 mm. The waist region 33 is in direct contact with the electrolyte and its decomposition products. Preferably, the optical fiber is made of Ge-As-Se-Te chalcogenide glass.

[0031] Preferably, the Ge-As-Se-Te chalcogenide glass is Ge20 As 20 Se 15 Te 45 Glass, which can transmit signals from 800 to 4000 cm. -1 The mid-infrared light signal.

[0032] The infrared fiber optic evanescent wave sensor in this embodiment is based on the infrared spectrum and can only detect substances in the 800-4000 cm⁻¹ range. -1 Selective absorption of infrared radiation of specific wavelengths causes transitions between vibrational and rotational energy levels in molecules.

[0033] The fabrication method of the above-mentioned infrared fiber optic evanescent wave sensor is as follows: Fiber drawing: High-purity Ge-As-Se-Te glass preforms are prepared using processes such as melt quenching, distillation purification, and secondary melting; then, the preforms are drawn into chalcogenide infrared optical fibers with a diameter of 500 μm using a high-precision fiber drawing tower.

[0034] Tapered structure fabrication: The original optical fiber is fixed on a high-precision tapered platform. The central area of ​​the optical fiber is softened by heating blocks. The stretching rate is controlled by programming a stepper motor to obtain a tapered micro / nano structure with a preset diameter waist area controlled at 100-200 μm and a length of 15 mm.

[0035] End face treatment and polishing: The two ends of the prepared tapered optical fiber are cut off, and the end face is finely polished using an optical fiber polishing device until the end face presents a scratch-free mirror state to ensure efficient coupling of the external infrared beam.

[0036] Battery model integration: Mounting holes matching the fiber optic diameter are drilled mechanically at pre-drilled locations on the battery model casing. The sensor is then slowly inserted through the battery along a pre-defined path, ensuring the waist-area sensing area is precisely aligned with the electrolyte site to be monitored.

[0037] Sealing and Encapsulation: Epoxy resin is used to hermetically seal the openings in the battery casing, securing the fiber optic leads to prevent displacement and ensuring complete isolation between the battery's interior and the external environment. After confirming no leakage through airtightness testing, sensor integration is completed.

[0038] Example 2: Infrared Fiber Evanescent Wave Sensing Device like Figure 3 As shown, this embodiment provides an infrared fiber optic evanescent wave sensing device, including an infrared light source, a ZnSe focusing lens, a chalcogenide cone-shaped micro / nano fiber optic sensor embedded in situ inside a battery, and an MCT liquid nitrogen-cooled detector.

[0039] This infrared fiber optic evanescent wave sensor is primarily used for in-situ real-time monitoring of electrolyte components and thermal decomposition products within lithium batteries. During actual detection, an infrared light source emits a mid-infrared beam, which is precisely focused by a ZnSe focusing lens guided by a beam-guiding coupling component, allowing the focused infrared light to be efficiently coupled into a chalcogenide tapered micro / nano fiber. As the light propagates within the fiber, an evanescent wave field is generated on its surface. The sensor, embedded in-situ within the battery, makes full contact with the electrolyte environment. Electrolyte molecules and decomposition products absorb the evanescent wave energy from the fiber surface. An MCT liquid nitrogen-cooled detector receives the transmitted light signal and converts it into an evanescent wave infrared absorption spectrum. By dynamically analyzing the intensity, position, and area of ​​characteristic absorption peaks in the spectrum, the system can obtain real-time structural information and concentration evolution patterns of key chemical components within the battery, providing fundamental data support for the early identification of thermal runaway.

[0040] By dynamically analyzing the area of ​​characteristic absorption peaks in the spectrum and retrieving their concentrations, the system can track the concentration evolution of solvent components inside the battery in real time and extract their dynamic consumption rate, thus providing core kinetic data support for the accurate identification of hidden chemical anomalies during the thermal runaway initiation phase.

[0041] Example 3: Application method of thermal runaway monitoring for lithium iron phosphate batteries This embodiment uses lithium-ion battery electrolytes containing DMC / EC (with lithium iron phosphate battery electrolyte as a typical example) as an example to provide specific usage methods for the infrared fiber optic evanescent wave sensor of Embodiment 1 and the infrared fiber optic evanescent wave sensing device of Embodiment 2 for monitoring the internal state of the battery. The specific usage methods are as follows: (1) Battery assembly and background collection Sensor preprocessing: Before formal testing, the background spectrum of the clean and dry chalcogenide tapered micro / nano fiber optic sensor is first collected to deduct the background noise of the environment and optical system.

[0042] In-situ embedding and encapsulation: The sensor is slowly embedded into the battery model along a preset path, ensuring that the waist area of ​​the sensor is completely immersed in the electrolyte.

[0043] Airtight sealing: Epoxy resin is used to airtightly seal the fiber optic lead-out end and mounting hole to ensure that there is no electrolyte leakage during battery testing and to maintain a stable internal chemical environment.

[0044] (2) Real-time spectral data acquisition optical path coupling and start-up: Place the battery model with the sensor loaded on the optical sensing platform (sensing device of Example 2), and adjust the position by the three-dimensional displacement stage to ensure that the infrared beam is coupled into the optical fiber with the highest intensity.

[0045] Parameter settings and scanning: Start the timing acquisition program and set the spectrometer resolution to 4 cm⁻¹. -1The scanning wavenumber range is 4000-800 cm⁻¹ -1 .

[0046] Thermal runaway-induced monitoring: The battery is heated at a preset heating rate to simulate the thermal runaway process. Sensors record in real time the infrared fingerprint spectral evolution of the electrolyte and its thermal decomposition products inside the battery.

[0047] Results analysis: By tracking the drift or intensity changes of characteristic functional group peaks through the acquired dynamic spectra, early chemical anomalies inside the battery can be identified. Figure 4 This section presents the evolution of the evanescent infrared absorption spectrum of the electrolyte during the entire thermal runaway process of the battery. We observed that as the temperature increases, the characteristic components in the electrolyte undergo significant decomposition, accompanied by the formation of new products. Specifically, at 842 cm⁻¹... -1 Gradual decomposition of LiPF6 was observed at 2350 cm⁻¹. -1 It is evident that the CO2 content generated by the decomposition of the electrolyte gradually increases, reaching a peak at 2923 cm⁻¹. -1 CH stretching vibrations can be observed, while the rest are EC / DMC, whose content gradually decreases due to thermal decomposition.

[0048] Example 4: Sensor Sensitivity Calibration and Quantitative Analysis Experiment To test the quantitative detection capability of the chalcogenide conical micro / nano fiber optic sensor described in this invention for the internal chemical components of a battery, this embodiment conducted sensitivity calibration experiments on key components of the electrolyte and typical thermal decomposition products: Sample preparation and background collection: Standard sample preparation: Prepare a series of standard solutions with concentration gradients of 10, 30, 50, 100, 300 and 500 mmol / L respectively.

[0049] Reactants (DMC and EC): Anhydrous acetonitrile was used as the solvent.

[0050] Background acquisition: Before formal testing, a background spectral scan is performed on the clean and dry sensor to eliminate environmental and system noise interference.

[0051] Characteristic peak selection and spectral acquisition: The infrared evanescent absorption spectra of samples of various concentrations are acquired sequentially using the sensing device of this invention (described in Example 1). Figure 5 As shown.

[0052] For quantitative evaluation, characteristic stretching vibration peaks were selected as response signals based on molecular fingerprint characteristics. Specifically, the peak at 1754 cm⁻¹ was chosen for dimethyl carbonate (DMC). -1 The nearby C=O stretching vibration peak, for ethylene carbonate (EC), is located at 1803 cm⁻¹. -1 Nearby C=O stretching vibration peak.

[0053] The results of sensitivity fitting and quantitative evaluation analysis show that the characteristic peak area (absorbance) of all sensors has a good positive correlation with the corresponding substance concentration, which conforms to the Lambert-Beer law. Figure 6 The sensitivities (slope S of the fitted line) of each component were obtained by linear fitting as follows: the sensitivity of DMC was 0.012 au / (mmol / L); the sensitivity of EC was 0.023 au / (mmol / L).

[0054] The coefficient of determination R for each fitted curve 2 All values ​​exceeded 0.99, demonstrating the high precision and reliability of the measurement system, which can capture effective chemical anomaly signals.

[0055] Example 5: Long-term stability assessment of the sensor in a complex electrolyte environment To ensure the sensor's practicality in scenarios such as energy storage stations, corrosion resistance and repeatability tests were conducted in this embodiment: Corrosion resistance test: The chalcogenide tapered micro / nano fiber optic sensor was completely immersed in an electrolyte containing 1M LiPF6 for one month.

[0056] Performance verification: Perform spectral analysis periodically and compare the test data from week 1 and week 4. Figure 7 Experimental results show that the background spectrum and characteristic peak signals have extremely high consistency, proving that Ge-As-Se-Te chalcogenide glasses have excellent chemical stability and resistance to electrolyte corrosion.

[0057] Repeatability testing: Figure 8 Multiple temperature-increasing cycle experiments were conducted on the same sensor, and the signal fluctuation range was controlled within the confidence interval, meeting the requirements for long-term online monitoring.

[0058] This invention first utilizes real-time acquired infrared evanescent wave spectral data, combined with calibrated component sensitivity curves, to construct a concentration-temperature correlation curve of electrolyte solvent components as the internal temperature of the battery evolves. Then, by analyzing and processing this curve, the dynamic consumption rate of the solvent components at different reaction stages is calculated. Finally, when the monitored solvent component consumption rate exceeds a preset dynamic threshold, the system determines that the internal chemical balance of the battery has been broken and has entered the thermal runaway initiation stage, thereby achieving accurate identification of hidden chemical anomalies and outputting corresponding early warning signals.

[0059] Example 6: Construction of a thermal runaway early warning model based on solvent consumption kinetics This embodiment provides a lithium battery thermal runaway early warning system, including: The detection module is used to collect infrared evanescent wave spectral data of electrolyte solvent components in situ in real time based on an infrared fiber optic evanescent wave sensor. The analysis and processing module is used to combine the electrolyte solvent component concentration calibration curve, invert the infrared evanescent wave spectral signal into the electrolyte solvent component concentration, and construct the concentration-temperature correlation curve of the electrolyte solvent component as the battery internal temperature evolves. The warning threshold module is used to analyze the concentration-temperature correlation curve of the electrolyte solvent components as the battery internal temperature evolves, and combine it with the dynamic consumption rate of the solvent components at different reaction stages to set the solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold. Specifically, the warning threshold includes a primary dynamic threshold k1 and a secondary threshold k2. The primary dynamic threshold k1 is the statistical upper limit value of the solvent dynamic consumption rate curve in the steady-state fluctuation stage, used to characterize the critical boundary when the monitoring signal deviates from the normal background noise and enters the early abnormal deviation stage. The secondary threshold k2 is the kinetic step starting point value of the solvent dynamic consumption rate curve from the stable evolution to the rapid acceleration stage, used to characterize the burst characteristic point of irreversible and violent thermal degradation of the electrolyte.

[0060] The determination module is used to determine that the internal chemical balance of the battery has been broken and the battery has entered the thermal runaway initiation stage when the consumption rate of the battery solvent component exceeds the preset dynamic threshold.

[0061] A corresponding early warning method is provided, including: S1 collects infrared evanescent wave spectral data of electrolyte solvent components in situ in real time based on the monitoring device described in claim 1; S2 combines the electrolyte solvent component concentration calibration curve to invert the infrared evanescent wave spectral signal into the electrolyte solvent component concentration, and constructs the concentration-temperature correlation curve of the electrolyte solvent component as the battery internal temperature evolves. S3 analyzes and processes the concentration-temperature correlation curve of electrolyte solvent components as the battery internal temperature evolves, and combines the dynamic consumption rate of solvent components at different reaction stages to set the solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold. When the S4 system detects that the consumption rate of the battery solvent components exceeds the preset dynamic threshold, it determines that the internal chemical balance of the battery has been broken and enters the thermal runaway initiation stage.

[0062] The following details how to transform the acquired raw infrared spectral signals into a quantifiable mathematical model for thermal runaway early warning, i.e., an early warning method.

[0063] (1) Preprocessing of raw spectral data First, in-situ infrared evanescent absorption spectra of the electrolyte collected at different temperature stages in Example 3 were obtained. Baseline correction was performed on the original spectra to eliminate baseline drift caused by refractive index fluctuations during the heating process; subsequently, difference spectral processing was performed using the spectrum at 30 °C as a reference. Figure 9Solvent characteristic peaks whose intensity decreases with increasing temperature can be clearly identified.

[0064] (2) Concentration inversion calculation Using the EC sensitivity coefficient S = 0.023 au / (mmol / L) calibrated in Example 4, according to the Lambert-Beer law...

[0065] Where A is absorbance (peak area), S is sensitivity slope, C is concentration, and b is intercept. (For 1803 cm⁻¹) -1 The characteristic peak area at a given point is inverted. The calculation formula is C=A-2.02 / 0.023, thus converting the absorbance signal into the residual solvent concentration C at different temperatures.

[0066] (3) Establishment of concentration-temperature correlation curve (CT curve) Plotting the battery's internal monitoring temperature T on the x-axis and the retrieved solvent residual concentration C on the y-axis as shown in the figure. Figure 10 The concentration-temperature correlation curve is shown.

[0067] This curve visually illustrates the decomposition stages of the solvent: Plateau phase (<120 ℃): The concentration decreases extremely slowly, indicating that the battery is in a stable operating state; Reaction initiation phase (120-200 ℃): The absolute value of the curve slope begins to increase, indicating that the electrolyte begins to undergo thermal decomposition; Violent reaction phase (>200 ℃): The curve shows a cliff-like drop, and the concentration drops sharply in a short period of time, corresponding to the irreversible thermal runaway exothermic chain reaction inside the battery.

[0068] (4) Extraction of consumption rate index Taking the first derivative of the above correlation curve yields the dynamic solvent consumption rate |dC / dT|, as shown below. Figure 11 This rate serves as a core kinetic indicator for evaluating the severity of internal side reactions within the battery. By capturing the step-like abrupt change in rate, it provides mathematical criteria for setting subsequent warning thresholds.

[0069] Example 7: Verification of the Timeliness of Early Warning Based on Dynamic Threshold k This embodiment is based on the dynamic model constructed in Embodiment 6, and the early warning performance of the system is verified in actual thermal abuse tests.

[0070] (1) Setting the warning threshold Scientific quantification based on electrolyte consumption kinetics ( Figure 11The first-level threshold k1, obtained through baseline fluctuations, represents the initial identification of chemical anomalies within the battery, marking the onset of irreversible early thermal degradation of the electrolyte. The second-level threshold k2, obtained through inflection point analysis of the kinetic curve, corresponds to the step-like transition from a gradual increase to an exponential surge in the |dC / dT| rate curve. Physically, this signifies the critical determination of irreversible thermal runaway, indicating that the solvent component has entered a stage of rapid, explosive heat generation and consumption. Through the coupled determination of these two dynamic thresholds, the system can accurately output graded early warning signals based on abrupt changes in the consumption rate at the molecular level before a significant temperature rise occurs on the battery surface.

[0071] Based on the consumption rate curve established in Example 6, a tiered early warning logic is set: Level 1 (Yellow) warning threshold k1: set at 1.5 mmol / (L) ℃). According to Figure 11 The curve, with the threshold slightly above the maximum fluctuation limit of the baseline plateau, is used to identify the initial shift of the solvent consumption rate away from the steady state.

[0072] Level 2 (Red) warning threshold k2: set at 3.5 mmol / (L) ℃). This threshold corresponds to Figure 11 The step start point (approximately 210 °C) at which the reaction enters a stage of rapid acceleration.

[0073] (2) Comparison with traditional monitoring technologies The initial temperature for the battery test was set at 30 ℃, and the ambient temperature rise rate was 10 ℃ / min.

[0074] Chemical anomaly budding stage (T=17 min): The internal temperature reaches 200 ℃. At this time, the |dC / dT| curve ends its low-level stable fluctuation range (~1.0) and shows an upward deviation trend; then, in the process of evolving towards 210 ℃, it rapidly breaks through the first-level warning threshold range.

[0075] Acceleration phase (T=18 min): Internal temperature reaches 210 ℃, and the reaction rate exceeds the secondary warning threshold of 3.5 mmol / (L). The temperature (℃) corresponds to a clear kinetic step point in the figure, indicating a precipitous consumption of solvent. After 250 ℃, the solvent consumption rate |dC / dT decreases significantly, indicating that the main components of the electrolyte (EC, DMC) have entered the reactant depletion stage. This completes the kinetic life cycle of the thermal runaway process from "start-up, acceleration" to "termination", achieving closed-loop monitoring of the entire process.

[0076] Traditional monitoring points: Under the same testing conditions, such as Figure 12As shown, the PT100 sensor attached to the battery surface needs to wait for internal heat to be conducted to the casing. When the battery surface temperature reaches the conventional safety threshold (80 °C), the test time has reached approximately 25 minutes.

[0077] Early warning lead time: This invention captures internal chemical signals in situ, providing an early warning time of approximately 8 minutes earlier than traditional methods in the first-level warning stage; and still leads traditional methods by approximately 7 minutes in the second-level emergency warning stage (i.e., when the violent internal reaction has just begun).

[0078] This embodiment demonstrates that by monitoring the solvent consumption rate in real time, the system can overcome the thermal conduction lag caused by battery thermal resistance and achieve "second-level" perception from the source of the chemical reaction, thus reserving a sufficient response window for the safety system.

[0079] Example 8: Verification Test of Thermal Runaway Early Warning for Soft-Pack Lithium-ion Batteries This embodiment uses a 10Ah aluminum-plastic film pouch lithium-ion battery as a typical example. The battery is placed in a heating device, and thermal abuse failure is induced at a constant heating rate of 10 °C / min. The experiment verifies the universality of this invention in practical packaging systems by embedding a GAST infrared fiber optic sensing unit in situ at the side seal. This scheme utilizes the principle of infrared evanescent waves to not only accurately capture the kinetic abrupt change in the solvent consumption rate |dC / dT|, but also to sensitively detect internal environmental disturbances caused by hidden gas generation in the early stages of thermal runaway in the pouch battery. Experimental results show that under a heating rate of 10 °C / min, when the system detects |dC / dT| reaching the first-level warning threshold, the battery surface temperature is only recorded as 40-60 °C, while the actual internal temperature has already climbed to nearly 200 °C. This indicates that this invention can provide a golden warning window of 5-8 minutes before the pouch battery experiences severe physical swelling or significant temperature rise, significantly improving the safety monitoring dimension of the pouch battery system.

Claims

1. A monitoring device for monitoring the internal structure of a battery, characterized in that: The device includes an infrared light source, a ZnSe focusing lens, an infrared fiber evanescent wave sensor, and a detector arranged in sequence. A beam of infrared light emitted by the infrared light source is focused and coupled into the infrared fiber evanescent wave sensor through the ZnSe lens, generating an evanescent wave at the interface between the fiber and the internal medium of the battery. The evanescent wave interacts with the electrolyte and its decomposition products coated near the waist region. The detector receives the optical signal from the fiber and converts it into an evanescent wave infrared absorption spectrum. The infrared fiber evanescent wave sensor includes a tapered chalcogenide micro / nano fiber optic sensing element that penetrates the battery casing and is embedded in situ inside the battery. The conical chalcogenide micro-nano light sensing element is a bare core structure, is prepared by using Ge-As-Se-Te chalcogenide glass, can transmit 800-4000 cm -1 of mid-infrared light, and the infrared transmission wave band covers 2-16 μm. The tapered chalcogenide micro / nano optical sensing element is drawn from optical fibers with a diameter of 150-1000 μm, including a transition region and a waist region, with the waist region diameter controlled between 100-200 μm. The conical sulfide-based micro / nano optical sensing element is embedded in situ inside the battery model and hermetically sealed.

2. The monitoring device according to claim 1, characterized in that: The Ge-As-Se-Te chalcogenide glass is Ge 20 As 20 Se 15 Te 45 The waist area length is 10-30 mm.

3. A method for early warning of battery thermal runaway based on solvent consumption kinetics, characterized in that: The method includes: S1 collects infrared evanescent wave spectral data of electrolyte solvent components in situ in real time based on the monitoring device described in claim 1; S2 combines the electrolyte solvent component concentration calibration curve to invert the infrared evanescent wave spectral signal into the electrolyte solvent component concentration, and constructs the concentration-temperature correlation curve of the electrolyte solvent component as the battery internal temperature evolves. S3 analyzes and processes the concentration-temperature correlation curve of electrolyte solvent components as the battery internal temperature evolves, and combines the dynamic consumption rate of solvent components at different reaction stages to set the solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold. When the S4 system detects that the consumption rate of the battery solvent components exceeds the preset dynamic threshold, it determines that the internal chemical balance of the battery has been broken and enters the thermal runaway initiation stage.

4. The method for early warning of battery thermal runaway according to claim 3, characterized in that: Based on the characteristic absorption peaks of the electrolyte solvent components in S2, and combined with the Lambert-Beer law, a linear quantitative relationship between absorbance and concentration was established to obtain the electrolyte solvent component concentration calibration curve.

5. The method for early warning of battery thermal runaway according to claim 3, characterized in that: In S3, the solvent dynamic consumption rate curve is obtained by taking the first derivative of the concentration-battery temperature correlation curve used to construct the solvent component. The solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold, is determined by analyzing the solvent dynamic consumption rate curve.

6. The method for early warning of battery thermal runaway according to claim 3, characterized in that: The warning threshold includes a primary dynamic threshold k1 and a secondary threshold k2. The primary dynamic threshold k1 is the statistical upper limit of the solvent dynamic consumption rate curve in the steady-state fluctuation stage, which is used to characterize the critical limit of the monitoring signal deviating from the normal background noise and entering the early abnormal deviation stage. The secondary threshold k2 is the kinetic step starting point value of the solvent dynamic consumption rate curve from the steady evolution to the rapid acceleration stage, which is used to characterize the burst characteristic point of irreversible and violent thermal degradation of the electrolyte.

7. The method for early warning of battery thermal runaway according to claim 3, characterized in that: Early warning thresholds are obtained based on the dynamic consumption rate analysis of electrolyte components ethylene carbonate or dimethyl carbonate.

8. The method for early warning of battery thermal runaway according to claim 3, characterized in that: The warning thresholds obtained based on the dynamic consumption rate analysis of the electrolyte component ethylene carbonate include a primary dynamic threshold and a secondary threshold. The primary dynamic threshold k1 is 1.5-2 mmol / (L) ℃); the secondary threshold k2 is 3.5-4 mmol / (L) ℃).

9. A lithium battery thermal runaway early warning system, characterized in that: include: The detection module is used to collect infrared evanescent wave spectral data of electrolyte solvent components in situ in real time based on the monitoring device described in claim 1; The analysis and processing module is used to combine the electrolyte solvent component concentration calibration curve, invert the infrared evanescent wave spectral signal into the electrolyte solvent component concentration, and construct the concentration-temperature correlation curve of the electrolyte solvent component as the battery internal temperature evolves. The warning threshold module is used to analyze and process the concentration-temperature correlation curve of the electrolyte solvent component as the internal temperature of the battery evolves, and combine the dynamic consumption rate of the solvent component at different reaction stages to set the solvent dynamic consumption rate |dC / dT|, i.e., the warning threshold. The determination module is used to determine that the internal chemical balance of the battery has been broken and the battery has entered the thermal runaway initiation stage when the consumption rate of the battery solvent component exceeds the preset dynamic threshold.

10. The application of the lithium battery thermal runaway early warning system according to claim 9 in new energy vehicles, safe energy storage power stations, and battery health status assessment.

Images