Lithium ion battery thermal runaway early warning method and system based on characteristic gas monitoring

Abstract

The invention discloses a lithium ion battery thermal runaway early warning method and system based on characteristic gas monitoring. The method comprises the following steps: S1, collecting the concentrations of C2H4, CH4 and CO characteristic gases in a lithium ion battery module in real time; s2, acquiring the current state of charge of the battery, namely acquiring an SOC value; s3, dynamically adjusting a characteristic gas concentration threshold according to the SOC value: (1) when SOC is less than or equal to 50%, adopting a reference threshold: C2H4 = 2.0 ppm; cH4 is equal to 10.0 ppm; cO is equal to 10.0 ppm; (2) when SOC is greater than 50%, calculating a dynamic threshold value according to a formula: C2H4 = 2 * e-0. 015 * (SOC-50) ppm; cH4 is equal to 10 * (1 to 0.008 * (SOC-50)) ppm; cO is equal to 10 * (1-0.006 * (SOC-50)) ppm; when it is monitored that the concentration of any characteristic gas exceeds a corresponding threshold value, a grading early warning signal is triggered.

Description

Technical Field

[0001] This invention relates to the field of battery safety monitoring technology, specifically to a method and system for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring. Background Technology

[0002] With the accelerated advancement of efforts to build new power systems and promote green energy transition, the importance of electrochemical energy storage facilities is becoming increasingly prominent. While the installed capacity of electrochemical energy storage power stations is growing rapidly, safety issues are also becoming increasingly apparent. Since 2017, nearly 40 safety accidents involving energy storage power stations have occurred both domestically and internationally, drawing significant public attention. Energy storage power stations suffer from large volumes of battery operation data, high management difficulty, and low operation and maintenance efficiency. Furthermore, batteries are prone to thermal runaway under abuse conditions, posing a high risk of combustion and explosion. Therefore, improving the health management and intelligent operation and maintenance of energy storage batteries has become a crucial measure to ensure the safe operation of energy storage power stations.

[0003] Thermal runaway in lithium-ion batteries is one of the most serious challenges in battery safety. Traditional early warning technologies mainly rely on temperature and voltage monitoring, but these suffer from response lag—by the time a temperature sensor triggers an alarm, the battery has often already entered an irreversible exothermic phase (such as SEI film decomposition, >80℃). In recent years, monitoring methods based on characteristic gases have become a research hotspot due to their potential for early warning, but existing technologies face three major technical bottlenecks: ① The contradiction between fixed threshold and dynamic gas generation characteristics: The battery gas generation behavior has a strong nonlinear relationship with the battery state of charge (SOC). Experimental data shows that when the SOC increases from 30% to 80%, The release rate increases exponentially (from 0.5 ppm / min to 4.2 ppm / min). If a fixed threshold is used, the warning time in high SOC scenarios is only 2-3 minutes, far below the 5-minute escape time window required by GB38031-2020. At low SOC, the same threshold leads to insufficient sensitivity, with a false alarm rate as high as 15%. ② Lack of multi-gas synergistic mechanisms: Existing technologies mostly use a single gas (such as CO) or a simple combination (…). However, the influence of electrolyte composition differences on characteristic gases was not considered. For example, CO accounts for over 60% during thermal runaway in LFP batteries, while NCM batteries release more during thermal runaway. Furthermore, electrolyte solvent vapors (such as DMC) can cause cross-interference and false alarms when humidity is >60%; ③ System-level response delay: In traditional solutions, sensors are mostly placed outside the battery module, resulting in a gas diffusion delay of 40-60 seconds. Additionally, fixed threshold algorithms cannot adapt to changes in gas production characteristics caused by battery aging. For example, after 500 cycles, the CO release threshold of the NCM battery decreases by 22%, and when the SOC increases from 50% to 100%,… With the release rate increased by 3.2 times, if a fixed threshold is still used, the warning time will be shortened to 3 minutes, which cannot meet the mandatory 5-minute requirement of GB 38031-2020. Therefore, there is an urgent need to develop a threshold algorithm based on dynamic adjustment of SOC. Summary of the Invention

[0004] The purpose of this invention is to develop a method for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring, which aims to overcome the technical bottlenecks faced by existing characteristic gas-based battery thermal runaway monitoring methods. It is applicable to energy storage power stations, electric vehicles, consumer electronics and other fields.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: This invention provides a method for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring, the method comprising the following steps: S1. Collect characteristic gas concentrations: Collect the concentrations of C2H4, CH4 and CO characteristic gases in the lithium-ion battery module in real time; S2. Obtain SOC value: Obtain the current state of charge of the battery, i.e., obtain the SOC value; S3. Dynamically adjust the characteristic gas concentration threshold based on the SOC value: ① When SOC ≤ 50%, the baseline threshold is used: C2H4=2.0ppm; CH4=10.0ppm; CO=10.0ppm; ②When SOC > 50%, the dynamic threshold is calculated using the formula: C2H4=2×e -0.015×(SOC-50) ppm; CH4=10×(1-0.008×(SOC-50)) ppm; CO=10×(1-0.006×(SOC-50)) ppm; When the concentration of any characteristic gas exceeds the corresponding threshold, a graded early warning signal is triggered.

[0006] Specifically, the core challenge of lithium-ion battery thermal runaway early warning lies in achieving early and accurate warnings under highly dynamic operating conditions. Currently, traditional fixed threshold methods cannot adapt to the significant impact of battery state of charge (SOC) changes on gas production behavior, resulting in delayed warnings in high SOC scenarios and frequent false alarms in low SOC scenarios. This invention proposes a lithium-ion battery thermal runaway early warning method based on characteristic gas monitoring. This method is a battery thermal runaway early warning method based on dynamic SOC threshold adjustment. By establishing a quantitative relationship between the characteristic gas release rate and SOC, an adaptive threshold calculation model is constructed, and combined with multi-sensor fusion technology, a reliable early warning more than 7 minutes before the onset of thermal runaway is achieved.

[0007] The technical solution of this invention first addresses the correlation between SOC and the release of characteristic gases at the mechanistic level. Through thermal runaway experiments on multiple sets of batteries with different SOCs (5-100%), combined with gas chromatography analysis, it was found that... C The release rate of CO shows a non-linear positive correlation with SOC; when SOC exceeds 50%, the electrolyte decomposition reaction intensifies. The release rate increases exponentially, while C Both CO and [the other variable] increase approximately linearly. Based on this, the present invention establishes a dynamic threshold calculation model, wherein... The threshold decreases exponentially with increasing SOC, C The threshold for SOC decreases linearly. Specifically, when SOC ≤ 50%, a baseline threshold is used ( =2.0ppm, C =10.0ppm, CO=10.0ppm); while when SOC>50%, The threshold is calculated using the formula 2×e (-0.015×(SOC-50)) ppm dynamic adjustment, C The thresholds for CO and CO decrease linearly with slopes of 0.8% and 0.6%, respectively. This dynamic threshold calculation model effectively solves the problem of warning delay caused by the accelerated gas production rate of high SOC batteries, while avoiding false triggering under low SOC conditions.

[0008] Specifically, the dynamic threshold model in this invention is constructed as follows: through several sets of thermal runaway experiments on batteries with different SOCs (5-100%), a gas production rate equation is established, and the equation is fitted using the least squares method to determine the gas production rate. C The threshold-SOC relationship curve between CO and SOC.

[0009] Furthermore, a method for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring: the SOC value is obtained in real time through the voltage-capacity integral method of the battery management system (BMS), with a sampling frequency ≥10Hz.

[0010] Furthermore, a method for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring: the graded early warning signals include a first-level early warning, a second-level early warning, and a third-level early warning; The trigger condition for a Level 1 warning is: if the concentration of a single characteristic gas exceeds the threshold, a Level 1 warning is triggered, and the ventilation system is activated to reduce the gas concentration. The triggering conditions for a Level 2 warning are: if the concentrations of two or more gases exceed the threshold simultaneously or the CO concentration exceeds 20.0 ppm, a Level 2 warning will be triggered, the charging and discharging circuit will be cut off, and an alarm will be issued. The triggering conditions for a Level III warning are: when the concentrations of all three characteristic gases exceed the threshold and the rate of temperature rise exceeds 1°C / second, it is determined that thermal runaway has entered an irreversible stage, and fire suppression is triggered.

[0011] The present invention also provides a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring: the early warning system includes the following settings: Several types of sensors are used to detect the concentrations of characteristic gases C2H4, CH4, and CO released from lithium-ion battery modules; The SOC detection module communicates with the battery management system to obtain the current SOC value of the battery. A dynamic threshold calculation unit, which contains a dynamic threshold calculation formula, is used to calculate the concentration threshold of each characteristic gas based on the SOC value. The early warning execution mechanism is linked with the controller, ventilation system, alarm system, and fire extinguishing system. It inputs the characteristic gas concentration data detected by the sensor and the threshold data calculated by the dynamic threshold calculation unit into the early warning execution mechanism to determine whether to trigger an early warning. If the triggering conditions corresponding to each level of early warning signal are met, the controller controls the ventilation system, alarm system, and fire extinguishing system to operate according to the respective level of early warning signal.

[0012] Furthermore, a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring: several sensors are respectively arranged at the pressure relief valve and gap of the lithium-ion battery module to detect the concentration of characteristic gases C2H4, CH4 and CO in the lithium-ion battery module.

[0013] Furthermore, a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring also includes a sensor located in the liquid collection tank area at the bottom of the battery. The sensor in this area is a composite sensor with a molecular sieve filter layer, used to filter non-characteristic gases to eliminate interference from electrolyte solvent vapors on the detection of characteristic gas concentration.

[0014] Furthermore, a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring: the sensor is also used to collect the battery's temperature rise rate, or the battery's temperature rise rate is collected by setting a temperature sensor.

[0015] Furthermore, a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring: the raw data collected by the sensor is input into the early warning execution mechanism after two-stage filtering; wherein, the two-stage filtering includes: firstly, using a moving average filter to eliminate pulse interference, and then applying a Kalman filtering algorithm for signal depth optimization.

[0016] Furthermore, a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring: If the early warning execution mechanism determines that the corresponding level of early warning triggering conditions are met, the early warning execution mechanism continuously monitors whether the characteristic gas concentration data input from the sensor drops below a threshold level within 10 seconds. If the characteristic gas concentration value exceeds the threshold for three consecutive monitoring periods during this period, the corresponding level of early warning response is triggered; otherwise, it is not triggered.

[0017] Specifically, the lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring provided by this invention employs a multi-channel sensor array (multiple sensors) and a high-precision SOC detection module working in tandem in its system architecture design. Sensors are positioned at key locations such as the pressure relief valve of the battery module, the gaps between battery cells, and the liquid collection tank at the bottom of the battery, forming a three-dimensional monitoring network to ensure no blind spots. The gas concentration data acquired by the sensors is denoised in real time using a Kalman filtering algorithm, and the dynamic threshold calculation cycle is controlled within 1 second to meet the requirements of rapid response. The early warning execution mechanism (early warning host) adopts a graded triggering mechanism: when the concentration of a single characteristic gas exceeds the threshold for 10 seconds, a first-level early warning is activated, and the ventilation system is activated to reduce the gas concentration; if two or more characteristic gases exceed the threshold at the same time or the CO concentration exceeds 20.0 ppm, a second-level early warning is triggered, the charging and discharging circuit is immediately cut off, and an audible and visual alarm is issued through the alarm system (the alarm system generally includes an alarm indicator light and a buzzer); when all three characteristic gases exceed the threshold and the temperature rise rate exceeds 1°C / second, it is determined that thermal runaway has entered an irreversible stage, the fire extinguishing system is automatically activated, and the power is immediately cut off. This graded early warning strategy of the present invention ensures safety while minimizing the impact of false alarms on system operation.

[0018] Furthermore, to further improve reliability, the early warning system of this invention also incorporates a false alarm suppression strategy. Specifically, to address interference caused by the volatilization of electrolyte solvents (such as DMC), the system uses a composite sensor with a molecular sieve filter layer to filter non-characteristic gases and combines this with a time window judgment logic: if the gas concentration momentarily exceeds the threshold but returns to normal levels within 10 seconds, it is determined to be environmental interference rather than actual thermal runaway; only when the gas concentration exceeds the threshold for 10 seconds continuously, and the gas concentration exceeds the threshold for 10 seconds in three consecutive monitoring sessions, is it confirmed as a valid early warning. Experimental data shows that this suppression strategy can control the false alarm rate to below 0.1%, significantly lower than the 5-10% false alarm level of traditional methods.

[0019] The beneficial effects of this invention are: The present invention provides a method for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring. C Combining CO as a characteristic gas, this system monitors changes in the concentration of this characteristic gas within the battery module in real time. By dynamically adjusting the gas concentration threshold in conjunction with the battery's state of charge (SOC), it achieves an early warning system that anticipates thermal runaway at least 7 minutes before it occurs. Specifically, this includes establishing a mathematical model of SOC and the release rate of the characteristic gas, and setting a baseline threshold (…). =2 ppm, C =10 ppm, CO=10ppm), when SOC>50%, the threshold is dynamically reduced according to the formula to improve sensitivity. This invention solves the problem of warning delay in high SOC scenarios by the traditional fixed threshold method. The warning time can be more than 40% earlier than the international standard, and the false alarm rate is less than 0.1%.

[0020] Thanks to the establishment of the early warning method and the design of the early warning system in this invention, the following excellent effects have been achieved: ① The early warning time is on average 40% earlier than that of the fixed threshold method. Especially when SOC≥80%, the early warning window can be extended to more than 9 minutes, far exceeding the 5-minute escape time window required by international standards; ② Through dynamic threshold adjustment and graded response mechanism, a balance between high sensitivity and low false alarm rate has been achieved; ③ The system has strong compatibility and can be adapted to various battery systems such as NCM and LFP through parameter calibration. In the future, it can also be extended to solid-state battery applications. Detailed Implementation

[0021] The technical solution of the present invention will be clearly and completely described below with reference to specific embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the present invention or its application or use. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] Example 1

[0023] A method for early warning of thermal runaway in lithium-ion batteries based on characteristic gas monitoring is provided, which includes the following specific steps: S1. Collect characteristic gas concentrations: Collect the concentrations of C2H4, CH4 and CO characteristic gases in the lithium-ion battery module in real time; S2. Obtain SOC value: Obtain the current state of charge (SOC value) of the battery in real time through the voltage-capacity integral method of the battery management system (BMS), with a sampling frequency ≥10Hz; S3. Dynamically adjust the characteristic gas concentration threshold based on the SOC value: ① When SOC ≤ 50%, the baseline threshold is used: C2H4=2.0ppm; CH4=10.0ppm; CO=10.0ppm; ②When SOC > 50%, the dynamic threshold is calculated using the formula: C2H4(threshold)=2×e -0.015×(SOC-50) ppm; CH4(threshold)=10×(1-0.008×(SOC-50)) ppm; CO(threshold)=10×(1-0.006×(SOC-50)) ppm; When the concentration of any characteristic gas exceeds the corresponding threshold, a graded warning signal (including Level 1 warning, Level 2 warning, and Level 3 warning) is triggered. The warning levels are as follows: Level 1: If the concentration of a single characteristic gas exceeds the threshold, a Level 1 warning is triggered, and the ventilation system is activated to reduce the gas concentration. Level 2: If the concentrations of two or more gases exceed the threshold simultaneously, or if the CO concentration exceeds 20.0 ppm, a Level 2 warning is triggered, the charging and discharging circuit is immediately cut off, and an alarm is issued. Level 3: If the concentrations of all three characteristic gases exceed the threshold and the temperature rise rate exceeds 1℃ / second, it is determined that thermal runaway has entered an irreversible stage, and fire suppression is triggered.

[0024] Example 2

[0025] Design a lithium-ion battery thermal runaway early warning system based on characteristic gas monitoring. The early warning system includes the following settings: Several types of sensors are respectively arranged at the pressure relief valve and the gap of the lithium-ion battery module to detect the concentration of characteristic gases C2H4, CH4 and CO in the lithium-ion battery module. A composite sensor with a molecular sieve filter layer is also set in the liquid collection tank area at the bottom of the battery to filter non-characteristic gases and eliminate the interference of electrolyte solvent vapor on the detection of characteristic gas concentration. The SOC detection module communicates with the battery management system (BMS) to obtain the current state of charge (SOC value) of the battery. A dynamic threshold calculation unit, which contains a dynamic threshold calculation formula, is used to calculate the concentration threshold of each characteristic gas based on the SOC value. The early warning execution mechanism is linked with the controller, ventilation system, alarm system, and fire extinguishing system. It inputs the raw gas concentration data collected from the sensors after two-stage filtering into the early warning execution mechanism (early warning host), and also inputs the threshold data calculated by the dynamic threshold calculation unit into the early warning execution mechanism to determine whether to trigger an early warning. If the early warning execution mechanism continuously meets the triggering conditions corresponding to each level of early warning signal, it continuously monitors whether the characteristic gas concentration data input from the sensors falls below the threshold level within 10 seconds. If the characteristic gas concentration value continuously exceeds the threshold during this period, and the monitoring results exceed the threshold for 10 seconds three consecutive times, the corresponding level of early warning response is triggered. The controller then controls the ventilation system, alarm system, and fire extinguishing system to operate according to the respective level of early warning signal; otherwise, no triggering occurs.

[0026] Application 1: The early warning method of Example 1 and the early warning system of Example 2 are applied to the power battery system of electric vehicles: taking a certain type of ternary lithium-ion power battery pack as the implementation object, the battery rated capacity is 78.5kWh, and it is composed of 96 individual cells forming an 8-parallel 12-series battery module. The system implementation process includes the following steps: (1) System initialization configuration: After the system is powered on, the battery type parameter library is loaded first, and the characteristic gas release kinetic parameters are automatically matched according to the battery model NCM622; then the sensor zero point calibration is performed, and 2.0 ppm of gas is introduced into the detection chamber. 10.0 ppm C With a standard gas of 10.0 ppm CO, the measurement error of each channel sensor is controlled within ±5%. The system synchronously initializes the dynamic threshold calculation module and loads the preset dynamic threshold-SOC relationship curve. (2) Sensor network deployment: A three-level sensor network is implemented inside the battery module: a MEMS infrared gas sensor (model TGS8100) is installed at the pressure relief valve at the top of the module for the main detection channel, with a detection accuracy of ±0.5ppm and a response time of <2 seconds; an electrochemical sensor (model SGP40) is arranged between each individual cell as an auxiliary detection node; a composite sensor with a molecular sieve filter layer is set in the liquid collection tank area at the bottom of the battery pack to eliminate interference from electrolyte solvent (DMC, etc.) vapor. All sensors are connected to the main control unit via CAN bus, and the sampling period is set to 500ms. (3) Real-time monitoring and data processing: When the system is running, the battery management system (BMS) calculates the SOC value in real time using the voltage-capacity integral method at a frequency of 10Hz and transmits it to the dynamic threshold calculation unit in the early warning host through a dedicated communication interface. The threshold of each gas is updated in real time according to the current SOC value. The original characteristic gas concentration data collected by the sensor is processed by two levels of filtering: first, the moving average filter is used to eliminate pulse interference, and then the Kalman filtering algorithm (process noise Q=0.01, observation noise R=0.05) is applied to optimize the signal depth. The processed data is input into the early warning execution mechanism (early warning host). (4) Warning Triggering and Execution: When the concentration of any gas exceeds the dynamic threshold, the warning execution mechanism starts a 30-second continuous judgment timer. If the concentration value continues to exceed the standard during this period, the corresponding level of warning response will be triggered: Level 1 warning (single gas exceeds the threshold) will control the centrifugal exhaust fan (ventilation system) on the top of the battery pack to force exhaust at a volume of 200 m³ / h; Level 2 warning (two or more gases exceed the threshold or CO concentration > 20.0 ppm) will directly cut off the main contactor through a hard wire signal, and at the same time send alarm information to the cloud monitoring platform through the vehicle T-BOX; Level 3 warning (three gases exceed the threshold and the temperature rise rate > 1℃ / s) will determine that thermal runaway has entered the irreversible stage, immediately activate the perfluorohexanone fire extinguishing system, and complete the high voltage power-off of the whole vehicle within 0.5 seconds; (5) System verification test: To verify the warning effect, a needle penetration triggering experiment was conducted at SOC=75%. The experimental data showed that: at t=0s, the needle penetration triggered a short circuit in the battery module, and at t=28s, the short circuit was triggered. The concentration reached 3.2 ppm (exceeding the dynamic threshold of 1.37 ppm, i.e., calculated according to the formula 2×e). (-0.015×(75-50)) The result was approximately 1.37 ppm. At t=65s, the CO concentration rose to 12.8 ppm (exceeding the threshold of 8.5 ppm). The system confirmed the level 2 warning and executed the circuit breaker operation at t=95s. The final thermal runaway occurred at t=635s, with a warning lead time of 9 minutes, which fully meets the requirements of GB 38031-2020 standard.

[0027] Application 2: Application in energy storage power station battery systems: An implementation was carried out for a 2.0MWh containerized energy storage system, consisting of 15 battery racks, each containing 24 lithium iron phosphate (LFP) battery modules. The implementation process was optimized based on Application 1 as follows: (1) Multi-module collaborative monitoring architecture: Based on the standard implementation method, a distributed detection strategy is added. Each battery rack is equipped with one area controller to collect data from six gas sensor nodes in the area in real time. The system adopts a "2 out of 3" voting mechanism: when two of the three adjacent modules trigger the first-level warning at the same time, the warning level of the area is automatically upgraded to the second level; when different area controllers detect the same warning mode, the whole-station emergency response is initiated. (2) System-level linkage control: The early warning system is deeply integrated with the power plant energy management system (EMS). When a level 2 early warning is triggered, the following linkage operations are automatically executed: the PCS is cut off and the grid connection switch is locked within 100ms; the emergency ventilation system is started (total air volume ≥ 5000m³ / h); the SOC of the affected battery cluster is forcibly limited to below 30% through the battery management system (BMS). At the same time, the system supports remote expert diagnosis mode, which can upload the real-time gas concentration spectrum to the monitoring center for trend analysis. (3) Extreme operating condition verification: An overcharge test (to SOC = 130%) was conducted under extreme conditions of ambient temperature 45℃ and relative humidity 80%. The test results showed that the system can effectively distinguish between actual thermal runaway gas production and environmental interference. When the CO concentration reached 15.0ppm (dynamic threshold 5.2ppm), the warning was accurately triggered without any false alarms. The concentration of DMC interfering gas passing through the molecular sieve filter layer was suppressed to below 1.0ppm, verifying the effectiveness of the anti-interference design.

[0028] The solution of this invention has been verified in electric vehicle and energy storage power station scenarios and has the potential for large-scale commercial promotion.

[0029] The above-described preferred embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of the invention. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the protection scope of the present invention.

Claims

1. A method for early warning of thermal runaway of a lithium ion battery based on characteristic gas monitoring, characterized in that, The method comprises the following steps: S1, real-time acquisition of C2H4, CH4 and CO characteristic gas concentration in lithium ion battery module; S2, obtaining the current state of charge of the battery, i.e. obtaining the SOC value; S3, dynamically adjusting the characteristic gas concentration threshold according to the SOC value: ① When SOC≤50%, the baseline threshold is adopted: C2H4=2.0ppm; CH4=10.0ppm; CO=10.0ppm; ② When SOC>50%, the dynamic threshold is calculated according to the formula: C2H4 = 2 x e -0.015×(SOC-50) ppm; CH4 = 10 x (1 - 0.008 x (SOC - 50)) ppm; CO=10×(1-0.006×(SOC-50)) ppm; When the concentration of any characteristic gas is detected to exceed the corresponding threshold, a hierarchical early warning signal is triggered.

2. The method of claim 1, wherein the method further comprises, The SOC value is obtained in real time by the voltage-capacity integration method of the battery management system, and the sampling frequency is ≥10Hz. 3.The lithium-ion battery thermal runaway early warning method based on characteristic gas monitoring of claim 1, wherein, The hierarchical early warning signal includes a first-level early warning, a second-level early warning and a third-level early warning; The first-level early warning trigger condition is that the concentration of a single characteristic gas exceeds the threshold, triggering the first-level early warning and starting the ventilation system to reduce the gas concentration. The second-level early warning trigger condition is that the concentrations of two or more gases exceed the threshold or the CO concentration exceeds 20.0ppm, triggering the second-level early warning, cutting off the charging and discharging circuit and issuing an alarm. The third-level early warning trigger condition is that the concentrations of three characteristic gases all exceed the threshold and the temperature rise rate exceeds 1℃ / s, determining that the thermal runaway has entered an irreversible stage and triggering the fire extinguishing.

4. The thermal runaway pre-alarming system for lithium-ion battery based on characteristic gas monitoring according to any one of claims 1-3, characterized in that, The early warning system comprises: A plurality of sensors for detecting the concentrations of C2H4, CH4 and CO characteristic gases released in the lithium ion battery module; An SOC detection module in communication connection with the battery management system for obtaining the current SOC value of the battery; A dynamic threshold calculation unit in which a dynamic threshold calculation formula is implanted for calculating the threshold values of each characteristic gas concentration according to the SOC value; An early warning execution mechanism in linkage with the controller, ventilation system, alarm system and fire extinguishing system, in which the characteristic gas concentration data detected by the sensors and the threshold data calculated by the dynamic threshold calculation unit are respectively input into the early warning execution mechanism for yes / no trigger early warning judgment, and if the judgment meets the trigger conditions corresponding to each level of early warning signal, the controller controls the ventilation system, alarm system and fire extinguishing system to operate according to each level of early warning signal.

5. The thermal runaway pre-alarm system for lithium-ion batteries based on characteristic gas monitoring according to claim 4, characterized in that, A plurality of sensors are arranged at the pressure relief valve and gap of the lithium ion battery module for detecting the concentrations of C2H4, CH4 and CO characteristic gases in the lithium ion battery module.

6. The thermal runaway pre-alarm system for lithium-ion batteries based on characteristic gas monitoring according to claim 5, characterized in that, A sensor is also arranged in the liquid collection tank area at the bottom of the battery, and the sensor arranged in this area is a composite sensor with a molecular sieve filter layer for filtering non-characteristic gases to eliminate the interference of electrolyte solvent vapor on the detection of characteristic gas concentration.

7. The thermal runaway pre-alarm system for lithium-ion battery based on characteristic gas monitoring according to claim 4, characterized in that, The sensor is also used to collect the temperature rise rate of the battery.

8. The thermal runaway pre-alarm system for lithium-ion battery based on characteristic gas monitoring according to claim 4, characterized in that, The raw data collected by the sensor is input into the early warning execution mechanism after two-stage filtering processing; The two-stage filtering processing includes first using sliding average filtering to eliminate pulse interference, and then applying Kalman filtering algorithm for signal optimization.

9. The thermal runaway pre-alarm system for lithium-ion battery based on characteristic gas monitoring according to claim 4, characterized in that, If the pre-warning execution mechanism judges that the pre-warning trigger condition of the corresponding level is met, the pre-warning execution mechanism continuously monitors whether the characteristic gas concentration data input by the sensor falls below the threshold level within 10 seconds, if the characteristic gas concentration value continuously exceeds the threshold value for 3 times during this period, the pre-warning response of the corresponding level is triggered, otherwise it is not triggered.