Multi-stage early warning method and system for thermal runaway of energy storage lithium ion battery

Abstract

The invention discloses an energy storage lithium ion battery thermal runaway multistage early warning method and system, and belongs to the technical field of lithium battery safety. The method comprises the steps that a three-dimensional closed-loop optical fiber monitoring network is constructed, distributed temperature sensing optical fibers are laid along the side face, the top face and the side face of a battery, and areas which are not smaller than 20 mm away from the center of a safety valve and not smaller than 15 mm away from a tab welding point are actively avoided; acquiring centimeter-level temperature field data through an optical fiber demodulator; and the processor calculates the temperature rise rate R and the temperature gradient G in real time based on the temperature field, and performs composite condition early warning. The triggering condition of the early warning is that when the temperature T of any measuring point reaches a preset threshold value, one of the temperature rise rate R and the temperature gradient G reaches the preset threshold value, and the proportion of all abnormal areas exceeding the threshold value in the total laying length of the optical fiber is not less than 30%, the early warning of the corresponding level is triggered. According to the method, local benign fluctuation can be effectively filtered, the false alarm rate is greatly reduced, and earlier and more accurate thermal runaway early prediction is realized.

Description

Technical Field

[0001] The present invention belongs to the field of lithium battery safety technology, and specifically relates to an early, high-precision, low-false-alarm-rate multi-level warning method for thermal runaway of large-capacity energy storage lithium-ion batteries, and an intelligent warning system for implementing the method. Background Art

[0002] As global energy transformation and climate change response become increasingly urgent, electrochemical energy storage systems, centered around lithium-ion batteries, are becoming a key enabler and are widely used in a variety of fields, including grid peak regulation, renewable energy integration, smart grids, electric transportation, and data centers. Lithium-ion batteries, with their high energy density, long cycle life, and increasing cost-effectiveness, have become the mainstream choice for large-scale energy storage.

[0003] However, the inherent safety of lithium-ion batteries remains a significant challenge to their large-scale application. Under the influence of extreme operating conditions, mechanical or electrical abuse, or internal defects, individual batteries (cells) may experience thermal runaway. Thermal runaway is a chain reaction process, typically accompanied by a sudden rise in internal battery temperature, gas production, expansion, and even rupture, which emits high-temperature, high-pressure gases and flames. This rapidly spreads to adjacent cells, causing catastrophic chain combustion and explosions, posing a significant threat to human life and property. Therefore, achieving early warning, precise location, and rapid intervention for thermal runaway in lithium-ion batteries has become a core focus of current research in battery technology and energy storage system safety.

[0004] Currently, battery thermal runaway warnings rely primarily on monitoring key battery parameters, including temperature, voltage, current, gas composition, and expansion force. Temperature is one of the most direct and important parameters for characterizing thermal runaway. Distributed fiber optic sensing technology, due to its inherent electrical insulation, electromagnetic interference resistance, high-temperature resistance, and unique advantages of continuous measurement along the line, is considered an ideal choice for battery temperature monitoring and has already found initial application in some battery management systems (BMS).

[0005] However, existing distributed optical fiber battery thermal runaway warning schemes still have significant limitations: First, fiber deployment lacks robustness. In many existing solutions, optical fibers are typically laid on the battery surface in a simple straight line or flat winding pattern. This layout does not fully consider the extreme physical environment when thermal runaway occurs. In particular, when the internal pressure of the battery is too high, causing the safety valve to open, the ejected high-temperature, high-pressure gas (reaching over 800°C and with a flow rate exceeding 200m / s) and flames can have a devastating impact on the optical fiber, easily causing it to fuse, resulting in a loss of monitoring capabilities at the most critical warning moment and causing the entire warning system to fail.

[0006] Second, the accuracy and anti-interference capabilities of the early warning criteria are poor. Most existing early warning systems still rely on fixed thresholds for single parameters (e.g., an alarm is triggered when the temperature exceeds a certain set value). This "point-based" judgment method is highly susceptible to factors such as local temperature fluctuations under normal operating conditions, changes in ambient temperature, heating caused by battery charging and discharging, and occasional poor contact, resulting in a large number of false alarms, with a false alarm rate as high as 35%. These frequent "crying wolf" false alarms not only reduce system credibility but also increase unnecessary operation and maintenance costs and the burden of emergency response. They are unable to effectively distinguish between real potential thermal runaway risks and harmless local hotspots.

[0007] Third, dynamic perception of fault development trends is insufficient. Existing systems often fail to fully utilize the vast amount of spatial temperature data provided by distributed optical fibers. Instead, they focus solely on the instantaneous value at a single point and lack in-depth analysis and utilization of the dynamic evolution of the temperature field over space and time (such as hotspot formation, diffusion rate, and direction). This makes it difficult to provide early warnings in the incipient stages of a fault, missing the optimal opportunity for intervention.

[0008] Therefore, there is an urgent need to develop a new battery thermal runaway warning method and system that can overcome the above-mentioned technical difficulties, so as to improve the reliability, accuracy and timeliness of the warning and fundamentally ensure the safe operation of the energy storage system. Summary of the Invention

[0009] The present invention aims to solve the problems in the prior art where distributed optical fiber sensors are easily damaged by thermal runaway jets due to improper deployment, resulting in monitoring failure, and where early warning criteria rely solely on a single-point threshold, resulting in a high false alarm rate. This provides a more reliable and accurate multi-level early warning method and system for thermal runaway of energy storage lithium-ion batteries.

[0010] To achieve the above objectives, the present invention provides a multi-level early warning method for thermal runaway of an energy storage lithium-ion battery, comprising the following steps: S1: Build a three-dimensional monitoring network: Use a distributed temperature-sensing optical fiber to lay along at least one side, top, and then another side of a single energy storage lithium-ion battery to form a closed-loop monitoring path; The monitoring path maintains a radial distance of not less than 20 mm from the center of the safety valve of the battery, and a distance of not less than 15 mm from the welding point of the battery tab of the battery; S2: Acquiring and analyzing temperature field data: using a fiber optic interrogator to acquire real-time temperature data of multiple measuring points on the distributed temperature-sensing optical fiber, and forming a battery surface temperature field with centimeter-level resolution; S3: Calculating dynamic evolution parameters: Based on the temperature field data, the temperature rise rate R of each measuring point and the temperature gradient G along the length of the optical fiber are calculated in real time; S4: Execute composite condition warning: When the following two conditions a) and b) are both satisfied, trigger a warning of the corresponding level: a) The real-time temperature T of at least one measuring point reaches the preset temperature threshold Tx, and the real-time temperature rise rate R reaches the preset rate threshold Rx, or the real-time temperature gradient G reaches the preset gradient threshold Gx; b) The proportion of the abnormal area where the temperature, temperature rise rate or temperature gradient exceeds their respective preset thresholds in the total laying length of the distributed temperature sensing optical fiber (2) is not less than 30%.

[0011] Preferably, in step S3, the temperature rise rate R is obtained by taking the time derivative dT / dt of the temperature of a single measuring point, and the temperature gradient G is obtained by taking the spatial derivative dT / dL of the temperature field data along the optical fiber laying path.

[0012] Preferably, it further includes: synchronously collecting the expansion force signal of the battery, the gas production concentration signal after the safety valve is opened, and the voltage signal of the battery as the basis for auxiliary warning judgment.

[0013] Preferably, the warning is a multi-level warning, including at least level one, level two and level three warnings; Among them, the preset temperature threshold Tx, rate threshold Rx and gradient threshold Gx are divided according to the warning level x, where x is 1, 2, 3, and T1 < T2 < T3, R1 ≤ R2 < R3, G1 < G2 < G3 are satisfied to distinguish different degrees of fault severity.

[0014] Preferably, between step S2 and step S3, it further includes: performing noise reduction processing on the real-time temperature data obtained by the optical fiber demodulator using moving average filtering, and the window width is 20 sampling points.

[0015] Correspondingly, the present invention also provides a multi-level warning system for thermal runaway of energy storage lithium-ion batteries, which executes the above method and includes: A distributed temperature sensing optical fiber: configured to be laid along at least one side surface, the top surface of a single energy storage lithium-ion battery, and then to another side surface to form a closed-loop monitoring path; the monitoring path maintains a radial distance of not less than 20 mm from the center of the safety valve of the battery, and a distance of not less than 15 mm from the welding point of the battery tab of the battery; A fiber optic demodulator: optically connected to the distributed temperature sensing optical fiber, configured to obtain centimeter-level real-time temperature data of multiple measuring points on the optical fiber based on the Brillouin scattering principle; A processor: electrically connected to the fiber optic demodulator, and the processor is configured to: Receive the real-time temperature data to construct the battery surface temperature field; Based on the temperature field data, the temperature rise rate R of each measuring point and the temperature gradient G along the length of the optical fiber are calculated in real time; When it is detected that the real-time temperature T and the real-time temperature rise rate R or the real-time temperature gradient G of at least one measuring point reach their respective preset thresholds, and the total length of all abnormal areas exceeding the corresponding thresholds accounts for no less than 30% of the total laying length of the optical fiber, an early warning signal is output.

[0016] Preferably, it also includes: a pressure sensor disposed between a metal fixture for holding the battery and the surface of the battery, for monitoring the expansion force of the battery; A gas sensor: connected to the top of the safety valve through a gas extraction pipe, used to monitor the gas concentration; Wherein, the pressure sensor and the gas sensor are both electrically connected to the processor.

[0017] Preferably, the distributed temperature-sensitive optical fiber comprises a fiber core, an aramid braided layer, a Teflon outer sheath and an internal metal reinforcing wire, and its overall temperature resistance is not less than 300°C.

[0018] Preferably, it also includes a plurality of polyimide high-temperature resistant insulating tapes for fixing the distributed temperature-sensitive optical fiber, the tapes fix the optical fiber at a spacing of 5 cm along the optical fiber laying path, and each fixing point can withstand a tensile force of not less than 50N.

[0019] Preferably, the total laying length L of the distributed temperature-sensing optical fiber on the surface of a single battery and the number N of surface temperature measurement points satisfy the relationship: L≈5cm×N, and N≥12.

[0020] Compared to existing technologies, this invention fundamentally addresses the fatal flaw in existing technologies of fiber optic sensors being easily blown by high-temperature, high-speed jets or arcs in the early stages of thermal runaway through an innovative three-dimensional closed-loop optical fiber layout ("side → top → side"), and by proactively avoiding specific safety distances from safety valves and battery tabs. This ensures the survivability of the monitoring system and the continuous reliability of data acquisition at the most critical moments. Building on this highly reliable data, this invention further abandons the traditional single-point threshold model and innovatively introduces a composite early warning criterion combining "physical quantity threshold" with "spatial abnormal area ratio." This criterion can effectively distinguish between dangerous heat sources with a spreading trend and benign, isolated local hotspots in terms of spatial dimensions, thereby accurately identifying the true signs of early thermal runaway and greatly reducing the high false alarm rate caused by fluctuations in normal operating conditions. It is this organic combination of highly reliable data acquisition and highly accurate dynamic analysis capabilities that enables the present invention to capture the subtle characteristics of abnormal heat accumulation and diffusion at the very early stages of a thermal runaway fault chain (such as abnormal expansion force and valve opening), even before macroscopic parameters such as voltage and temperature undergo drastic changes. This allows it to issue an early warning more than 200 seconds earlier than traditional methods, thus gaining extremely valuable response time for taking emergency intervention measures.

[0021] In summary, the present invention, through a systematic solution design, comprehensively solves the three major industry pain points of sensor survivability, warning accuracy, and response timeliness, providing a more comprehensive and reliable technical guarantee for the intrinsic safety of energy storage systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of an instrument and equipment according to an embodiment of the present invention; Figure 2 Schematic diagram of the arrangement of a battery thermal runaway test system according to an embodiment of the present invention; Figure 3 Schematic diagram of the arrangement of optical fibers on the surface of a battery in an embodiment of the present invention; Figure 4 It is the time sequence diagram of multi-dimensional signals and temperature cross-section diagram at different moments during the battery thermal runaway process; Figure 5 This is a spatiotemporal evolution diagram of multiple physical quantities measured by distributed optical fiber sensors during the thermal runaway experiment. Specifically, it includes: Figure 5 (a) is a 2D pseudo-color image of the battery surface temperature field along the fiber position and time evolution; Figure 5 (b) is the abnormal expansion force (t Abn )、Safety valve open(t v ) and internal short circuit (t ISC ) One-dimensional distribution curves of temperature along the fiber position at three key moments; Figure 5 (c) is a 2D pseudo-color image of the temperature rise rate field along the fiber position and time evolution; Figure 5 (d) is the abnormal expansion force (t Abn )、Safety valve open(t v ) and internal short circuit (t ISC ) One-dimensional distribution curves of temperature rise rate at three key moments; Figure 5 (e) is a 2D pseudo-color image of the temperature gradient field along the fiber position and time evolution; Figure 5 (f) is the expansion force abnormality (t Abn )、Safety valve open(t v ) and internal short circuit (t ISC ) One-dimensional distribution curves of temperature gradient at three key moments.

[0023] Figure 6 This is a spatiotemporal evolution diagram of multiple physical quantities measured under normal 1C rate cycling conditions (a total of 3 cycles), used for comparison with thermal runaway conditions. Specifically including: Figure 6 (a) is the 2D spatiotemporal distribution diagram of the battery surface temperature field; Figure 6 (b) is the 2D spatiotemporal distribution diagram of the temperature rise rate field; Figure 6 (c) in the figure is the 2D spatiotemporal distribution diagram of the temperature gradient field.

[0024] Figure 7 The three-dimensional scatter plot shows the correlation between the three key indicators of the maximum battery surface temperature, maximum temperature rise rate and maximum temperature gradient under two different working conditions.

[0025] Figure 7 (a) shows the distribution of data points in three normal cycles; Figure 7 (b) shows the distribution of data points during the thermal abuse (500W heating) process until thermal runaway, which is used to visually compare the evolution paths of warning parameters under normal and abnormal conditions. In the figure: 1-fiber optic demodulator, 2-temperature sensing optical fiber, 3-processor, 4-paperless recorder, 5-thermocouple, 6-24V power supply, 7-pressure sensor, 8-gas sensor, 9-exhaust pipe, 10-video recorder, 11-battery, 12-safety valve, 13-battery tab, 14-heating plate, 15-insulation cotton, 16-metal clamp, 17-stud bolt, 18-high temperature resistant test bench. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention. Example 1

[0027] This example describes a multi-stage early warning method for thermal runaway in lithium-ion energy storage batteries. This method aims to achieve early, high-precision, and low-false-alarm monitoring and early warning of thermal runaway risks in lithium-ion batteries. This method relies not only on advanced sensing technology but also on an innovative monitoring network layout and integrated intelligent early warning criteria. The following provides an in-depth and detailed description of each step of this method, along with accompanying figures.

[0028] At the initial stage of the early warning process, a highly reliable three-dimensional monitoring network is constructed (step S1). This step forms the physical foundation for all subsequent data analysis and early warning decisions. Its core design objective is to ensure comprehensive and accurate temperature measurement while also guaranteeing the sensor's survivability in extreme thermal runaway events. In specific implementation, a distributed temperature-sensing optical fiber 2 with centimeter-level spatial resolution is selected. For example, a fiber with an outer diameter of 2.4 mm and specially reinforced can be used. Its structure includes a fiber core, an aramid braid, a Teflon outer sheath, and internal metal reinforcement wires, ensuring an overall temperature resistance of no less than 300°C. This ensures that the fiber maintains structural and functional integrity even when the battery surface temperature rises sharply. This distributed temperature-sensing optical fiber 2 is laid along a specific path along the surface of a single energy storage lithium-ion battery 11. The energy storage lithium-ion battery 11 can be a large-capacity prismatic battery, such as the 280 Ah lithium iron phosphate battery used for experimental verification in this description. The laying path is not a simple two-dimensional planar winding, but rather follows an optimized three-dimensional closed-loop topology of "side → top → other side". Specifically, the distributed temperature-sensing optical fiber 2 can start from one main side surface of the battery 11, extend upward to cover its top surface, and then extend downward to the other main side surface of the battery 11, ultimately forming a closed-loop monitoring path that can monitor the key heat-generating areas of the battery in a three-dimensional and no-dead-angle manner. This closed-loop design also provides the possibility for signal integrity verification.

[0029] In the process of building this three-dimensional monitoring network, the key innovation of this embodiment is reflected in the precise layout requirements for proactively avoiding two known high-risk areas. Figure 2 and Figure 3 The hint, Figure 2 shows the layout of the entire test system, and Figure 3This more clearly depicts the specific location of the optical fiber on the battery surface. The monitoring path of the distributed temperature-sensing optical fiber 2 is strictly regulated, requiring a radial distance of at least 20 mm from the geometric center of the safety valve 12 at the top of the battery 11. This distance is determined based on an in-depth analysis of the physical process of thermal runaway. When the pressure inside the battery reaches the design threshold due to side reactions, the safety valve 12 will rupture or open, instantly ejecting a mixture of high-temperature, high-pressure gas, electrolyte vapor, and flames at temperatures reaching up to 800°C and velocities exceeding 200 m / s. The core impact zone is extremely destructive. Over 90% of fiber optic sensor failures in existing technology are caused by routing the sensor too close to this area. By enforcing a safety distance of at least 20 mm, the distributed temperature-sensing optical fiber 2 is ensured to be located at the edge of the jet's impact, greatly improving its survival probability in a thermal runaway event. Similarly, the monitoring path must also maintain a distance of at least 15 mm from the weld point of the battery tab 13 of the battery 11. The battery tab 13 is a key component connecting the battery's internal and external circuits. In the event of an external short circuit or a worsening internal short circuit, this area can become an area of ​​intense Joule heat, potentially generating a high-temperature arc and causing thermal damage to nearby materials. Maintaining a safe distance of at least 15mm can effectively prevent damage from this localized extreme heat.

[0030] At the same time, in order to ensure that the distributed temperature-sensing optical fiber 2 always maintains close and stable thermal contact with the battery surface during the long-term operation of the battery due to environmental vibration or its own thermal expansion and contraction, thereby ensuring the accuracy and response speed of temperature measurement, this embodiment uses polyimide high-temperature resistant insulating tape for segmented fixation. Optionally, the fixed spacing can be set to 5cm. This tape is not only resistant to high temperatures, but also has excellent mechanical properties. For example, its thickness can be less than or equal to 0.1mm, and each fixed point can withstand a tensile force of not less than 50N, ensuring that it can still provide a firm and reliable fixing effect at an ambient temperature of up to 200°C. In addition, in order to ensure the precision of monitoring, the total laying length L of the distributed temperature-sensing optical fiber 2 on the surface of a single battery 11 and the number N of surface temperature measurement points should satisfy a specific relationship, for example, L≈5cm×N, and N≥12, to ensure that for less than 5cm 2 Effective capture of the early thermal runaway core area.

[0031] After the high-reliability three-dimensional monitoring network is constructed, step S2 is performed: acquiring and analyzing temperature field data. This step is accomplished using a fiber optic interrogator 1 optically connected to the distributed temperature-sensing optical fiber 2. The fiber optic interrogator 1 is the core of the entire sensing system, and its operating principle is preferably based on Brillouin scattering. A laser within the fiber optic interrogator 1 periodically emits a probe light pulse into the distributed temperature-sensing optical fiber 2. As the light pulse propagates forward through the fiber, it interacts with the optical fiber medium, generating various backscattered light patterns. The center frequency of the Brillouin scattered light exhibits a slight frequency shift relative to the incident light, known as the Brillouin frequency shift. The magnitude of this frequency shift is linearly proportional to the temperature and strain at the location of the optical fiber. Using a highly sensitive photodetector and a high-speed signal processing unit, the fiber optic interrogator 1 receives and analyzes this backscattered light, which carries both position information (determined by the light pulse's time of flight) and temperature information. This allows for real-time, distributed analysis of precise temperature values ​​at multiple measurement points along the optical fiber. The system used in this method has a spatial resolution of ±0.05m and a temperature accuracy of ±0.05°C, thus forming an extremely detailed dynamic "heat map" of the battery surface, providing a high-density data foundation for subsequent dynamic evolution analysis.

[0032] After obtaining the high-resolution temperature field data, step S3 is immediately executed, i.e., calculating the dynamic evolution parameters. This step is usually completed by a processor 3 electrically connected to the fiber optic demodulator 1. After the processor 3 receives the continuous temperature field data stream transmitted by the fiber optic demodulator 1, its built-in algorithm program does not only focus on the static absolute value of the temperature, but further deeply explores the dynamic change characteristics of the temperature field in the time and space dimensions, because these dynamic characteristics can better reveal the trend of heat accumulation and diffusion, and are the key to distinguishing normal operating temperature rise from abnormal thermal runaway precursors. Specifically, the processor 3 will calculate two core dynamic parameters for each measuring point in real time: the temperature rise rate R and the temperature gradient G.

[0033] The rate of temperature rise, R, is calculated by taking the first-order temporal derivative of the temperature data at a single measurement point at successive time points. Its mathematical expression can be simplified to: R = dT / dt. This parameter intuitively reflects the speed of temperature change at a fixed location. A continuously and rapidly increasing positive value, R, typically indicates dangerous heat accumulation at that location.

[0034] The temperature gradient G is obtained by taking the spatial first-order derivative of the temperature field data along the path of the distributed temperature-sensing optical fiber 2 at the same moment in time. Its mathematical expression can be simplified to: G = dT / dL. This parameter characterizes the degree of temperature difference between physically adjacent measuring points. An abnormally large G value indicates the formation of a local hotspot, from which heat is spreading to the surrounding area.

[0035] Before calculating these two parameters, as a preferred technical solution, the processor 3 may first perform a denoising process on the original real-time temperature data obtained from the fiber optic demodulator 1. For example, a moving average filtering algorithm can be adopted, and a moving window with a width of 20 sampling points can be set, and the data within the window is arithmetically averaged. This can effectively smooth out the random noise introduced by electromagnetic interference or the system itself, thereby significantly improving the stability and accuracy of the subsequent calculation results of the temperature rise rate R and the temperature gradient G.

[0036] Finally, the most core decision-making link in this method flow is step S4, that is, to execute the composite condition warning. This is the fundamental difference between the present invention and the prior art in the warning logic, aiming to essentially solve the problem of the false alarm rate of up to 35% in the traditional single-point threshold warning method. The processor 3 internally presets multi-level warning criteria corresponding to different degrees of fault severity. Optionally, the warning levels can be divided into at least level one, level two, and level three. Each level of warning corresponds to a set of precise thresholds, including the temperature threshold Tx, the rate threshold Rx, and the gradient threshold Gx. These thresholds are not set arbitrarily, but are obtained based on a large amount of experimental data on battery thermal abuse and the analysis and statistics of the thermal failure mechanism.

[0037] For example, in this embodiment, based on the analysis of Figure 4 , Figure 5 , Figure 6 and Figure 7 the experimental data shown, a set of optimized three-level warning thresholds can be determined: The three-level temperature thresholds are T1 = 100°C, T2 = 160°C, and T3 = 240°C respectively; The three-level temperature rise rate thresholds are R1 = 0.2°C / s, R2 = 0.2°C / s, and R3 = 1.0°C / s respectively; The three-level temperature gradient thresholds are G1 = 4°C / cm, G2 = 6°C / cm, and G3 = 8°C / cm respectively.

[0038] Among them, the magnitude relationship of the thresholds satisfies T1 < T2 < T3, R1 ≤ R2 < R3, G1 < G2 < G3, which is used to distinguish different stages such as the budding, development, and near out-of-control of the fault.

[0039] The triggering of the warning is not a simple single-point overrun, but requires simultaneously satisfying the following two logical AND conditions a) and b): Condition a) is the triggering condition of the "point", which ensures the sensitivity of the warning. That is, among all the measurement points monitored by the entire distributed temperature sensing optical fiber 2, as long as the real-time temperature T of at least one measurement point reaches the preset temperature threshold Tx of the corresponding level, and at the same time its real-time temperature rise rate R reaches the preset rate threshold Rx of the corresponding level, or its real-time temperature gradient G reaches the preset gradient threshold Gx of the corresponding level, this condition is satisfied.

[0040] However, satisfying condition a) alone is not enough to trigger an immediate warning. Condition b), the "surface" confirmation condition, must also be satisfied. This is the key to this method's ability to suppress false alarms. Condition b) requires calculating the percentage of a spatial region.

[0041] First, define the fiber area covered by all measurement points whose real-time values ​​exceed their respective preset thresholds as abnormal areas. Then, calculate the proportion of the total length of these abnormal areas to the total length of the distributed temperature-sensing optical fiber 2 laid on the battery surface. The calculation method is: Abnormal area = Σ(abnormal area fiber length × 0.05m); Total area = total length of single cell surface laying L×0.05m; Proportion = abnormal area / total area.

[0042] Condition b) is only met if the calculated "abnormal spatial area ratio" is no less than 30%. Processor 3 will finally confirm and trigger the corresponding level of warning only if both conditions a) and b) are met.

[0043] In addition, in order to objectively and comprehensively verify and compare the early warning effect of the method of the present invention, traditional monitoring methods will be used and relevant data will be recorded during the experiment. Figure 1 and Figure 2 As shown, a traditional thermocouple 5 can be arranged at the center of the front and rear surfaces of the battery 11 to measure the temperature of a single point. The signal of the thermocouple 5 will be connected to a paperless recorder 4, which is responsible for collecting and displaying the temperature data of the thermocouple 5 in real time, and can store it for subsequent analysis. At the same time, in order to visually observe the entire process of thermal runaway, a video recorder 10 will be set up at the experimental site to record the entire process of the test device. The data collected by the thermocouple 5, paperless recorder 4 and video recorder 10 will be used as a reference benchmark to evaluate the improvement of the timeliness and accuracy of the warning method of the present invention.

[0044] To illustrate the effectiveness of this method more vividly, see Figures 4 to 7 The experiment used a 280Ah lithium iron phosphate battery as the object and heated it with a 500W heating plate 14 to simulate the thermal abuse process. Figure 4 The upper part shows the battery voltage, back center temperature T back and the heating plate temperature T heatThe lower half shows the changes in expansion force, carbon monoxide, and carbon dioxide concentrations over time. Vertical dashed lines mark several key fault characteristic moments in the figure. In the figure, the battery experiences abnormal expansion force, valve opening, internal short circuit, and thermal runaway at 1022 seconds, 2420 seconds, 3266 seconds, and 3343 seconds, respectively.

[0045] Now, we apply the method of the present invention to analyze the same process. Figure 5 , which shows in detail the spatiotemporal evolution of the multi-physics field monitored by the method of the present invention during the thermal runaway process. Specifically: Figure 5 (a) and Figure 5 (c) in the figure shows the two-dimensional temporal and spatial distribution of the temperature field and the temperature rise rate field measured by the distributed temperature sensing optical fiber 2; and Figure 5 (b) and Figure 5 (d) shows the one-dimensional cross-sectional distribution curves of temperature and temperature rise rate along the fiber position at the three key moments of abnormal expansion force, opening of safety valve and internal short circuit. It is crucial that Figure 5 (e) and Figure 5 (f) in the figure clearly reveals the evolution of another core parameter, the temperature gradient. Figure 5 As can be seen from the time-space diagram (e) in Figure 3, as thermal runaway approaches, obvious local high-temperature areas appear on the battery surface, resulting in a sharp temperature gradient. Figure 5 The one-dimensional cross-sectional view in (f) shows quantitatively that in the case of internal short circuit (t ISC ), the maximum temperature gradient soared to 14.03°C / cm, which strongly proved the formation of localized dangerous hot spots, which is the key feature that the early warning method of the present invention can accurately capture.

[0046] At the moment of abnormal expansion force, that is, 1022 seconds, the highest temperature, maximum temperature rise rate and maximum temperature gradient monitored by the optical fiber reached 102.8℃, 0.23℃ / s and 5.56℃ / cm respectively.

[0047] At the moment of pressure release, i.e. 2420 seconds, these three values ​​reached 187.2°C, -0.58°C / s and 9.66°C / cm respectively.

[0048] At the moment of internal short circuit, i.e. 3266 seconds, these three values ​​soared to 275.1°C, 7.31°C / s and 14.03°C / cm.

[0049] Combining this data with our three-level warning threshold and the 30% proportion rule, we can conclude: The first-level warning (T≥100℃ and R≥0.2℃ / s or T≥100℃ and G≥4℃ / cm, and the proportion is ≥30%) was triggered at about 787 seconds, which was about 235 seconds earlier than the 1022 seconds when the expansion force was abnormal.

[0050] The second-level warning (T≥160℃ and R≥0.2℃ / s or T≥100℃ and G≥6℃ / cm, and the proportion is ≥30%) was triggered at about 1876 seconds, which was about 544 seconds earlier than the 2420 seconds for valve opening.

[0051] The third-level warning (T≥240℃ and R≥1.0℃ / s or T≥100℃ and G≥8℃ / cm, and the proportion is ≥30%) was triggered at about 3001 seconds, which was about 265 seconds earlier than the 3266 seconds of the internal short circuit.

[0052] This fully demonstrates the great advantage of the present invention in terms of early warning timeliness.

[0053] In order to demonstrate the excellent anti-interference ability of the present invention, we compare its performance under normal working conditions with the above thermal runaway process. Figure 6 , This figure shows the monitoring results of the same battery under normal 1C rate cycle conditions (a total of 3 cycles). Figure 6 (a) shows the temporal and spatial distribution of the temperature field. Figure 6 (b) shows the temporal and spatial distribution of the temperature rise rate field, and Figure 6 (c) shows the spatiotemporal distribution of the temperature gradient field. It can be seen that although in normal cycles, the maximum temperature of the battery can reach 97.7°C, the maximum temperature rise rate reaches 0.99°C / s, and the maximum temperature gradient reaches 3.3°C / cm, some parameter values ​​may exceed the threshold of the first-level warning at certain moments, meeting condition a). However, since these temperature rises are holistic and uniform, and do not form localized abnormal hot zones with a spreading trend, the proportion of its spatial abnormal areas is always far below 30%, which does not meet condition b). This can be seen from Figure 7 (a) and Figure 7 This is more intuitively confirmed by comparing the three-dimensional scatter plot in (b): Figure 7 In (a), the data points under normal cycle are densely distributed in a low-risk area, while Figure 7 The data points during the thermal runaway process (b) show a clear evolutionary path toward the high-risk region. Therefore, the method of the present invention does not generate any false positives during the entire normal cycle, demonstrating its excellent anti-interference ability.

[0054] The multi-level early warning method for thermal runaway in lithium-ion energy storage batteries, described in detail in this embodiment, ensures a robust sensor data foundation through proactive, avoidant design in physical deployment. It also enhances insight into the nature of faults through in-depth exploration of multi-dimensional dynamic evolution parameters of the temperature field. Finally, through its innovative "point-to-surface" composite early warning criteria, it achieves a high degree of consistency in both warning sensitivity and reliability. This constitutes a complete and effective battery safety early prevention technology solution with significant technical effectiveness, significantly enhancing the safety of energy storage systems in practical applications. Example 2

[0055] This embodiment provides a multi-level early warning system for thermal runaway of energy storage lithium-ion batteries. This system is the specific physical implementation and engineering carrier of the method described in Example 1. As an integrated hardware and software platform, the system is designed to solidify the various steps and algorithms of the aforementioned method into a stable and reliable physical product to ensure that it can operate efficiently and autonomously in various practical energy storage application scenarios, such as large container-type energy storage power stations, electric vehicle battery packs, or data center backup power systems. Figure 1 and Figure 2 The system's core components include a distributed temperature-sensing optical fiber 2, a fiber optic interrogator 1, and a processor 3 serving as the central processing unit. Multiple auxiliary sensors can also be integrated. These components, through precise physical installation positioning and clear signal connections, work together to achieve real-time, multi-dimensional monitoring and graded early warning of thermal runaway risks in energy storage lithium-ion batteries 11.

[0056] First of all, the perception basis and core sensing element of the system is a specially designed and deployed distributed temperature-sensing optical fiber 2. In terms of physical installation, the optical fiber is configured to fit tightly against the surface of a single energy storage lithium-ion battery 11. Its installation position and laying path strictly follow the three-dimensional monitoring network construction principle detailed in Example 1, that is, along at least one side of the battery 11, through its top surface, and then to the other side, forming a closed-loop monitoring path. In the specific installation and construction, high-precision positioning tools must be used to ensure that the routing path of the optical fiber maintains a radial distance of not less than 20 mm from the geometric center point of the safety valve 12 of the battery 11, and also maintains a distance of not less than 15 mm from the welding point of the battery tab 13 of the battery 11. This installation layout that avoids specific hazardous areas is an inherent requirement of the design of this system and is the key to ensuring that it can still work normally under extreme conditions. To ensure the performance and lifespan of the sensor, the distributed temperature-sensing optical fiber 2 itself is optionally manufactured from high-temperature-resistant, high-strength materials, with an overall temperature tolerance of no less than 300°C. Its physical structure may include, in sequence, a central optical fiber core, an aramid braided layer for reinforcement, a Teflon outer sheath for chemical and physical protection, and built-in metal reinforcement wires for enhancing tensile strength. In addition, the system also includes multiple polyimide high-temperature-resistant insulating tapes for firmly securing the distributed temperature-sensing optical fiber 2 to the battery surface. These tapes are configured to securely secure the optical fiber to the battery surface at intervals of, for example, 5 cm along the fiber's laying path, thereby ensuring efficient and rapid heat conduction and long-term structural stability. Each fixing point can withstand a tensile force of no less than 50N.

[0057] Next, the signal demodulation and data conversion unit of the system is a fiber optic demodulator 1. In terms of physical connection, an optical port of the fiber optic demodulator 1 is optically connected to one end of the distributed temperature-sensing optical fiber 2 through a standard fiber optic jumper and connector, thereby forming a complete and closed optical path. The fiber optic demodulator 1 is configured to work based on the Brillouin scattering principle. Its core function is to emit detection light pulses into the optical fiber and to receive and demodulate the backscattered light signals returned from various points along the optical fiber with high sensitivity. It integrates a precise light source, an optical circulator, a photodetector and a high-speed signal processing module, which can accurately convert the slight drift of the light wave frequency into temperature readings at all measuring points along the line and output them in the form of digital signals. These data are reliably electrically connected and transmitted in real time to the system's central processing unit processor 3 through its data communication interface, such as an Ethernet port or an RS485 interface.

[0058] The core control and decision-making unit of the system is a processor 3. Figure 1In the laboratory environment shown, processor 3 can be a data logging computer with computing and data logging capabilities. However, in broader industrial or commercial applications, it is more likely to be a dedicated, highly reliable industrial control computer (IPC), a programmable logic controller (PLC), or a high-performance embedded system motherboard customized for specific needs. In the system connection, processor 3 is electrically connected to the fiber optic interrogator 1 to continuously receive its output, a real-time temperature data stream with centimeter-level resolution.

[0059] The processor 3 is the core embodiment of the intelligence of the entire system, and the software program solidified or running inside it is precisely configured to execute all the core algorithms of the method described in the first embodiment.

[0060] Specifically, the processor 3 is first configured to receive continuous real-time temperature data and build a battery surface dynamic temperature field model in its memory that can be refreshed in real time.

[0061] It is then configured to calculate two key dynamic evolution parameters, the temperature rise rate R and the temperature gradient G along the length of the optical fiber, in real time and point by point based on this temperature field data.

[0062] Finally, and most crucially, processor 3 is configured to execute the present invention's "point-to-surface" composite conditional early warning logic. It continuously compares the real-time temperature T, real-time temperature rise rate R, and real-time temperature gradient G at each measuring point with pre-programmed multi-level thresholds (Tx, Rx, Gx). It also simultaneously calculates, in real time, the ratio of the total length of all abnormal regions exceeding the corresponding thresholds to the total fiber installation length. Once a parameter value at a measuring point reaches a pre-set threshold (satisfying condition a), and the calculated abnormal region ratio simultaneously reaches or exceeds a pre-set ratio, such as 30% (satisfying condition b), processor 3 immediately generates a warning signal of the corresponding level at its output port. This warning signal can take various forms. For example, it can be a level flip of one or more digital I / O ports, which is used to directly drive a relay to disconnect the charging circuit or activate a fire sprinkler. It can also be an audible and visual alarm. Or it can be a network data packet that follows a specific protocol (such as Modbus TCP or CAN). Detailed warning information (including warning level, trigger time, abnormal area location, etc.) is sent via industrial Ethernet or CAN bus to the central monitoring system or battery management system (BMS) of the entire energy storage power station for higher-level decision-making and response.

[0063] In order to more comprehensively demonstrate the composition of this system and its application in the experimental environment, Figure 2The three-dimensional structure of the entire test device is depicted. The entire device is placed on a sturdy high-temperature resistant test bench 18. The battery 11 under test is clamped by a metal clamp 16 together with a heating plate 14 used to simulate thermal abuse, and a certain preload is applied and maintained by a stud bolt 17. In order to reduce the loss of heat to the outside world and more accurately simulate the thermal insulation environment of the battery in the module or battery pack, the battery 11 and the heating plate 14 are wrapped with a layer of thermal insulation cotton 15. This complete experimental device construction ensures the validity and repeatability of the experimental data, and provides a verification platform for the system of the present invention that is close to the actual working conditions.

[0064] In order to further enhance the system's early warning capabilities and fault diagnosis dimensions, the system of this embodiment can also selectively integrate other types of sensors to form a multi-dimensional information fusion platform. Figure 2 As shown, the system can include a pressure sensor 7, carefully positioned between the metal clamp 16 holding the battery 11 and the surface of the battery 11. This allows for non-invasive, real-time monitoring of changes in the macroscopic expansion force of the battery due to gas production from irreversible side reactions within the battery. The system can also include a gas sensor 8, specifically positioned via a gas extraction tube 9, with its sampling port precisely positioned directly above the safety valve 12. This allows for immediate detection of the presence and concentration changes of escaping characteristic gases (such as carbon monoxide (CO) and hydrogen (H2), etc.) upon opening the safety valve 12. Regarding system connectivity, these additional pressure sensors 7 and gas sensors 8 are electrically connected to the processor 3 via their own signal cables. The processor 3 software is configured to collect and analyze their signals, using them as a powerful supplement and cross-correlation with the fiber optic temperature data, enabling multi-information fusion and judgment, thereby further improving the accuracy of early warning decisions and the precision of fault diagnosis. The entire system, including the optical fiber interrogator 1, the processor 3 and all attached sensors, can be powered by a stable and reliable DC power supply 6, for example, a 24V industrial-grade switching power supply.

[0065] The multi-level early warning system for thermal runaway of energy storage lithium-ion batteries described in detail in this embodiment completely and reliably implements all the innovative methods proposed in Example 1 through the precise configuration of its various components, unique physical installation locations, and clearly defined signal connection relationships. From the fiber optic sensor network layout with high survivability at the front end, to the high-precision photoelectric demodulation unit at the mid-end, to the processor with highly intelligent composite algorithms at the back end, the entire system forms a seamless closed loop from physical perception to intelligent decision-making. It can serve as an independent and complete technical solution to stably and reliably perform early warning tasks for thermal runaway in various complex practical applications, and can effectively respond to the severe safety challenges of energy storage systems. The excellent early warning capabilities and effective suppression of false alarms demonstrated in the experiment fully demonstrate its advanced nature, practical value, and huge market application potential as an overall technical solution.

Claims

1. A multi-level early warning method for thermal runaway of energy storage lithium-ion batteries, characterized in that: It includes the following steps: S1: Construct a three-dimensional monitoring network: A distributed temperature sensing optical fiber (2) is used and laid along at least one side surface, the top surface of a single energy storage lithium-ion battery (11), and then to another side surface to form a closed-loop monitoring path; wherein, the monitoring path maintains a radial distance of not less than 20 mm from the center of the safety valve (12) of the battery (11), and maintains a distance of not less than 15 mm from the welding point of the battery tab (13) of the battery (11); S2: Obtain and analyze temperature field data: A fiber optic demodulator (1) is used to obtain the real-time temperature data of multiple measuring points on the distributed temperature sensing optical fiber (2) to form a battery surface temperature field with centimeter-level resolution; S3: Calculate dynamic evolution parameters: Based on the temperature field data, the temperature rise rate R of each measuring point and the temperature gradient G along the length direction of the optical fiber are calculated in real time; S4: Execute composite condition warning: When the following two conditions a) and b) are simultaneously satisfied, a warning of the corresponding level is triggered: a) The real-time temperature T of at least one measuring point reaches the preset temperature threshold Tx, and at the same time the real-time temperature rise rate R reaches the preset rate threshold Rx, or the real-time temperature gradient G reaches the preset gradient threshold Gx; b) The proportion of the abnormal area where the temperature, temperature rise rate or temperature gradient exceeds their respective preset thresholds in the total laying length of the distributed temperature sensing optical fiber (2) is not less than 30%.

2. The multi-level early warning method for thermal runaway of an energy storage lithium-ion battery according to claim 1, characterized in that: In step S3, the temperature rise rate R is obtained by taking the time derivative dT / dt of the temperature of a single measuring point, and the temperature gradient G is obtained by taking the spatial derivative dT / dL of the temperature field data along the optical fiber laying path.

3. The multi-level early warning method for thermal runaway of an energy storage lithium-ion battery according to claim 1, characterized in that: It further includes: Simultaneously collecting the expansion force signal of the battery (11), the gas production concentration signal after the safety valve (12) is opened, and the voltage signal of the battery (11) as the basis for auxiliary warning judgment.

4. The multi-level early warning method for thermal runaway of an energy storage lithium-ion battery according to claim 1, characterized in that: The warning is a multi-level warning, including at least level 1, level 2 and level 3 warnings; wherein, the preset temperature threshold Tx, rate threshold Rx and gradient threshold Gx are divided according to the warning level x, x is 1, 2, 3, and satisfy T1<T2<T3, R1≤R2<R3, G1<G2<G3 to distinguish different degrees of fault severity.

5. The multi-level early warning method for thermal runaway of an energy storage lithium-ion battery according to claim 1, characterized in that: Between step S2 and step S3, it further includes: Denoising the real-time temperature data obtained by the fiber optic demodulator (1) by using moving average filtering, and the window width is 20 sampling points.

6. A multi-level early warning system for thermal runaway of a lithium-ion battery for energy storage, which implements the method described in any one of claims 1 to 5, characterized in that: It includes: A distributed temperature sensing optical fiber (2): Configured to be laid along at least one side surface, the top surface of a single energy storage lithium-ion battery (11), and then to another side surface to form a closed-loop monitoring path; the monitoring path maintains a radial distance of not less than 20 mm from the center of the safety valve (12) of the battery (11), and maintains a distance of not less than 15 mm from the welding point of the battery tab (13) of the battery (11); A fiber optic demodulator (1): Optically connected to the distributed temperature sensing optical fiber (2), configured to obtain centimeter-level real-time temperature data of multiple measuring points on the optical fiber based on the Brillouin scattering principle; A processor (3) electrically connected to the optical fiber demodulator (1), the processor being configured to: Receiving the real-time temperature data to construct a battery surface temperature field; Based on the temperature field data, the temperature rise rate R of each measuring point and the temperature gradient G along the length of the optical fiber are calculated in real time; When it is detected that the real-time temperature T and the real-time temperature rise rate R or the real-time temperature gradient G of at least one measuring point reach their respective preset thresholds, and the total length of all abnormal areas exceeding the corresponding thresholds accounts for no less than 30% of the total laying length of the optical fiber, an early warning signal is output.

7. The multi-level early warning system for thermal runaway of an energy storage lithium-ion battery according to claim 6, characterized in that: Also includes: a pressure sensor (7) disposed between a metal clamp (16) for clamping the battery (11) and the surface of the battery (11), for monitoring the expansion force of the battery; A gas sensor (8) connected to the top of the safety valve (12) via a gas extraction pipe (9) for monitoring the gas concentration; Wherein, the pressure sensor (7) and the gas sensor (8) are both electrically connected to the processor (3).

8. The multi-level early warning system for thermal runaway of an energy storage lithium-ion battery according to claim 6, characterized in that: The distributed temperature-sensitive optical fiber (2) comprises a fiber core, an aramid braided layer, a Teflon outer sheath, and an internal metal reinforcement wire, and its overall temperature tolerance is not less than 300°C.

9. The multi-level early warning system for thermal runaway of an energy storage lithium-ion battery according to claim 6, characterized in that: It also includes a plurality of polyimide high-temperature resistant insulating tapes for fixing the distributed temperature-sensing optical fiber (2), wherein the tapes fix the optical fiber at a spacing of 5 cm along the optical fiber laying path, and each fixing point can withstand a tensile force of not less than 50N.

10. The multi-level early warning system for thermal runaway of an energy storage lithium-ion battery according to claim 6, characterized in that: The total laying length L of the distributed temperature-sensing optical fiber (2) on the surface of a single battery (11) and the number N of surface temperature measurement points satisfy the relationship: L≈5cm×N, and N≥12.

Images