Lithium battery thermal runaway prediction and multistage response system based on AI intelligent evaluation

Abstract

The invention discloses a lithium battery thermal runaway prediction and multistage response system based on AI intelligent evaluation, and the system comprises a multi-sensing monitoring module, an AI prediction evaluation module, a thermal runaway risk evaluation module, a safety response control module, and a safety execution module. Through fusion of multi-dimensional sensing data and an artificial intelligence algorithm, real-time prediction and evaluation of the thermal runaway risk of the battery are realized, and graded safety response measures are actively triggered at the early stage of thermal runaway. And aiming at different risk levels, the system automatically executes corresponding multi-level protection actions. The system also has online learning and model adaptive optimization capabilities, and can continuously improve the prediction accuracy and response efficiency along with the change of the state of the battery. And the thermal runaway early warning sensitivity and the response timeliness are remarkably improved, and the method has an intelligent level and engineering practicability and can be widely applied to the field of lithium battery safety management of electric vehicle battery packs, energy storage power stations and the like.

Description

Technical Field

[0001] This invention relates to the field of lithium battery technology, specifically to a lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment. Background Technology

[0002] Lithium-ion batteries are widely used in electric vehicles, energy storage power stations, and portable electronic devices due to their high energy density and good cycle life. However, under extreme conditions such as overcharging, over-discharging, internal short circuits, and high temperatures, lithium batteries may experience thermal runaway, triggering a violent exothermic reaction that causes a rapid rise in battery temperature and the release of flammable and toxic gases (such as hydrogen fluoride (HF) and carbon monoxide (CO). In severe cases, thermal runaway can cause the battery to catch fire and explode, endangering personal and equipment safety.

[0003] Existing battery management systems (BMS) and safety devices primarily employ passive protection mechanisms: typically, power-off, cooling, or fire suppression measures are only triggered after the battery temperature reaches a fixed threshold. This simple fixed-threshold control method suffers from response lag; when action is taken only when the battery temperature approaches the critical point, it is often too late to prevent thermal runaway in time. Furthermore, traditional devices lack effective means to handle the large amounts of toxic gases generated during thermal runaway, generally relying solely on natural venting through pressure relief vents in the battery casing, which fails to prevent the spread of harmful gases and causes secondary harm to the environment and personnel.

[0004] To improve the thermal runaway protection of lithium batteries, some studies have attempted to introduce advanced prediction and control technologies. For example, some research utilizes digital twin technology to simulate the internal state of the battery in real time, or employs machine learning / deep learning models to perform big data-driven risk prediction of the battery state. However, these cutting-edge technologies often suffer from problems such as complex models, high computational load, and strong dependence on historical data, resulting in insufficient stability and reliability in real-time applications in actual battery systems. Furthermore, current battery thermal management solutions lack robust active intervention strategies (such as variable cooling intensity and multi-level response), and lack the ability to automatically switch control logic based on the severity of the fault.

[0005] Therefore, there is an urgent need for a method that integrates reliable artificial intelligence prediction and evaluation with rule-based control on the basis of existing experimental platforms to proactively control the multi-level safety response to lithium battery thermal runaway. This method should achieve early warning of thermal runaway, rapid intervention and cooling, suppression of heat propagation, and treatment of toxic gases, while ensuring system stability and real-time performance, thereby addressing the issues of response lag and insufficient protection in existing technologies. Summary of the Invention

[0006] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0007] 1. Technical problems to be solved:

[0008] To address the aforementioned issues of delayed response and insufficient protection, this invention is proposed.

[0009] Therefore, the purpose of this invention is to provide a lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment. In response to the above-mentioned problems of untimely thermal runaway response and incomplete protection, this system is provided to achieve early prediction and warning of lithium battery thermal runaway and multi-level active safety protection.

[0010] 2. Technical Solution:

[0011] To address the aforementioned technical problems, according to one aspect of the present invention, the present invention provides the following technical solution:

[0012] It includes a multi-sensor monitoring module, an AI prediction and assessment module, a thermal runaway risk assessment module, a safety response and control module, and a safety execution module:

[0013] The multi-sensor monitoring module is used to monitor the operating parameters of the lithium-ion battery, including battery temperature, voltage, current, and gas concentration in the battery compartment environment, and to acquire the parameter data in real time.

[0014] The AI ​​prediction and evaluation module is used to perform intelligent analysis and thermal runaway trend prediction based on the operating parameter data, and output the evaluation index of battery thermal runaway risk or the short-term predicted value of battery temperature.

[0015] The thermal runaway risk assessment module is used to identify abnormal operating conditions of the battery based on the operating parameters and the prediction assessment results. When the battery temperature and temperature rise rate exceed the preset threshold, an abnormal voltage drop trend occurs, or the AI ​​assessment index exceeds the risk threshold, the thermal runaway risk level is determined.

[0016] The safety response control module is used to trigger a corresponding multi-level safety response strategy based on the thermal runaway risk level, and generate control commands to execute predetermined safety response measures, including cooling, power regulation, gas emission treatment, and circuit isolation.

[0017] The safety execution module includes a cooling device, a gas suppression device, a power-off device, and a fire extinguishing device, which are used to cool the battery, divert and purify harmful gases, cut off circuits, and extinguish fires at different response levels, respectively, in order to suppress the further spread of thermal runaway.

[0018] As a preferred embodiment of the lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment of the present invention, the AI ​​prediction and assessment module uses a recurrent neural network-based algorithm to perform time series analysis on the multi-source monitoring data and outputs the risk probability or temperature trend prediction value of lithium battery thermal runaway; wherein the recurrent neural network includes a long short-term memory (LSTM) network or a lightweight deep neural network, which is used to train and learn on continuously monitored data to extract time series features.

[0019] As a preferred embodiment of the lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment of the present invention, the thermal runaway risk assessment module is preset with multiple judgment conditions, including a battery temperature threshold, a temperature rise rate threshold, a voltage drop magnitude threshold, and an AI risk score threshold. When the battery temperature or temperature rise rate is detected to exceed the safety threshold, or the battery voltage shows an abnormal downward trend, or the risk score output by the AI ​​model exceeds the risk score threshold, the corresponding thermal runaway risk level is determined and a risk alarm signal is output.

[0020] As a preferred embodiment of the AI-based intelligent assessment-based lithium battery thermal runaway prediction and multi-level response system of the present invention, the multi-level safety response strategy includes at least three response levels: Level 1 warning, Level 2 intervention, and Level 3 emergency protection. Specifically, in the Level 1 warning state, the system issues an audible and visual alarm and activates the conventional cooling mode of the cooling device to control the battery temperature rise trend; in the Level 2 intervention state, the system reduces the battery power output to a predetermined safety level, activates the enhanced cooling mode, and turns on the gas suppression device to extract and purify the gas in the battery compartment; in the Level 3 emergency protection state, the system disconnects the battery from the external circuit and triggers the fire extinguishing device to extinguish the fire in the battery.

[0021] As a preferred embodiment of the AI-based intelligent assessment lithium battery thermal runaway prediction and multi-level response system of the present invention, the AI ​​prediction and assessment module has online learning and model adaptive optimization functions. By incrementally training and updating the model parameters based on the battery data collected during operation, the artificial intelligence model can continuously improve the accuracy of thermal runaway risk prediction as the battery ages and operating conditions change.

[0022] As a preferred embodiment of the AI-based intelligent assessment-based lithium battery thermal runaway prediction and multi-level response system of the present invention, the cooling device has a conventional cooling mode and an emergency cooling mode; when the battery temperature exceeds a first preset threshold, the conventional cooling mode is activated to cool down; when the temperature continues to rise and exceeds a second preset threshold, it automatically switches to the emergency cooling mode to enhance the cooling intensity. The conventional cooling dissipates heat through a fan or liquid cooling circulation, while the emergency cooling rapidly suppresses the rise in battery temperature by coolant injection or phase change material heat absorption.

[0023] As a preferred embodiment of the AI-based intelligent assessment-based lithium battery thermal runaway prediction and multi-level response system of the present invention, the gas suppression device includes a battery compartment depressurization channel and a gas filtration and purification module connected to the channel. When the battery experiences thermal runaway and releases harmful gases, the channel opens under the control of the safety response control module, guiding the gas in the battery compartment into the filtration and purification module for treatment. The filtration and purification module contains activated carbon adsorbent and catalytic conversion medium, used to adsorb toxic gases such as HF and CO and convert them into harmless products, thereby reducing the concentration of harmful gases and the risk of combustion.

[0024] 3. Beneficial effects:

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] This AI-based intelligent assessment-based lithium battery thermal runaway prediction and multi-level response system:

[0027] 1. Early and Accurate Warning: By combining AI intelligent models with multi-sensor monitoring, this invention achieves more accurate capture and risk assessment of early signs of thermal runaway in lithium batteries. Compared to traditional single-threshold monitoring, this invention can identify subtle danger signals such as abnormal temperature rises and abnormal voltage fluctuations, and issue early warning commands based on AI prediction results, greatly shortening the lag time of thermal runaway response;

[0028] 2. Comprehensive Tiered Response: This invention establishes a comprehensive multi-tiered safety response strategy, from initial early warning (alarm prompts and temperature reduction) to mid-term intervention (power limiting, enhanced cooling, and harmful gas purification) and finally to emergency response (power outage isolation, automatic fire suppression). Each level of measure is progressive and interconnected, automatically switching to the optimal control strategy based on the severity of the accident and AI risk assessment results. This tiered prevention and control system effectively avoids premature or delayed responses, improving the targeting and effectiveness of safety measures.

[0029] 3. Integrated Management of Heat and Gas: This invention incorporates a specialized cooling mode switching and gas suppression linkage mechanism, simultaneously addressing both overheating and harmful gas issues during thermal runaway. Dynamic switching of cooling methods maintains efficient heat dissipation under normal operating conditions and rapidly cools the battery in emergencies, preventing the spread of thermal runaway. Gas diversion and filtration promptly extract and purify leaked flammable and toxic gases, significantly reducing the risk of fire and poisoning. The combination of these two methods comprehensively controls the chain reactions of hazards caused by battery thermal runaway.

[0030] 4. Intelligent, Efficient, and Continuously Optimized: This invention introduces AI algorithms into battery safety monitoring and organically combines them with rule-based control, improving the sensitivity of risk identification while ensuring system real-time performance and reliability. The AI ​​model used is small in scale and computationally efficient, avoiding reliance on massive historical data and high-performance hardware by focusing on short-term predictions, and its performance can be continuously optimized using online learning mechanisms. While ensuring rapid response, it achieves intelligent identification and early intervention of thermal runaway risks, possessing good engineering practicality and easy integration into existing battery safety management platforms. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0032] Figure 1 This is a flowchart of a lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment, according to the present invention. Detailed Implementation

[0033] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0034] This invention is described in detail with reference to the schematic diagrams. When describing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0035] The orientation or positional relationship indicated in the terminology is based on the orientation or positional relationship shown in the accompanying drawings and is only for the convenience of describing the invention and simplifying the description, and is not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention.

[0036] The term "connection method" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0037] The embodiments of the present invention will now be described in further detail with reference to the accompanying drawings.

[0038] This invention provides an overall structural schematic diagram of an embodiment of a lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment, comprising:

[0039] Please see Figure 1 This embodiment of a lithium battery thermal runaway prediction and multi-level response system based on AI intelligent assessment includes a multi-sensor monitoring module, an AI prediction and assessment module, a thermal runaway risk assessment module, a safety response control module, and a safety execution module.

[0040] The multi-sensor monitoring module is used to monitor the operating parameters of lithium-ion batteries, including battery temperature, voltage, current, and gas concentration in the battery compartment environment, and to acquire parameter data in real time.

[0041] The AI ​​prediction and assessment module is used to perform intelligent analysis and predict thermal runaway trends based on operating parameter data, and output assessment indicators of battery thermal runaway risk or short-term predicted values ​​of battery temperature.

[0042] The thermal runaway risk assessment module is used to identify abnormal operating conditions of the battery based on operating parameters and predictive assessment results. When the battery temperature and temperature rise rate exceed the preset threshold, an abnormal voltage drop trend occurs, or the AI ​​assessment index exceeds the risk threshold, the thermal runaway risk level is determined.

[0043] The safety response control module is used to trigger corresponding multi-level safety response strategies based on the thermal runaway risk level, and generate control commands to execute predetermined safety response measures, including cooling, power regulation, gas emission treatment, and circuit isolation.

[0044] The safety execution module includes a cooling device, a gas suppression device, a power-off device, and a fire extinguishing device, which are used to cool the battery, divert and purify harmful gases, cut off circuits, and extinguish fires at different response levels, respectively, in order to suppress the further spread of thermal runaway.

[0045] It is worth noting that, specifically, the AI ​​prediction and evaluation module uses a recurrent neural network-based algorithm to perform time series analysis on multi-source monitoring data and output the risk probability or temperature trend prediction value of lithium battery thermal runaway; the recurrent neural network includes a long short-term memory (LSTM) network or a lightweight deep neural network, which is used to train and learn on continuously monitored data to extract time series features.

[0046] Next, specifically, the thermal runaway risk assessment module has several preset judgment conditions, including battery temperature threshold, temperature rise rate threshold, voltage drop magnitude threshold, and AI risk score threshold. When the battery temperature or temperature rise rate exceeds the safety threshold, or the battery voltage shows an abnormal downward trend, or the risk score output by the AI ​​model exceeds the risk score threshold, the corresponding thermal runaway risk level is determined and a risk alarm signal is output.

[0047] Specifically, the multi-level safety response strategy includes at least three response levels: Level 1 warning, Level 2 intervention, and Level 3 emergency protection. In Level 1 warning, the system issues an audible and visual alarm and activates the conventional cooling mode of the cooling device to control the battery temperature rise. In Level 2 intervention, the system reduces the battery power output to a predetermined safety level, activates the enhanced cooling mode, and turns on the gas suppression device to extract and purify the gas in the battery compartment. In Level 3 emergency protection, the system disconnects the battery from the external circuit and triggers the fire extinguishing device to extinguish the fire on the battery.

[0048] Furthermore, specifically, the AI ​​prediction and evaluation module has online learning and model adaptive optimization functions. By incrementally training and updating the model parameters based on the battery data collected during operation, the artificial intelligence model can continuously improve the accuracy of predicting thermal runaway risk as the battery ages and operating conditions change.

[0049] Specifically, the cooling device has a regular cooling mode and an emergency cooling mode. When the battery temperature exceeds the first preset threshold, the regular cooling mode is activated to cool down the battery. When the temperature continues to rise and exceeds the second preset threshold, the device automatically switches to the emergency cooling mode to enhance the cooling intensity. The regular cooling mode dissipates heat through a fan or liquid cooling circulation, while the emergency cooling mode quickly suppresses the rise in battery temperature by injecting coolant or absorbing heat through phase change materials.

[0050] Next, specifically, the gas suppression device includes a battery compartment depressurization channel and a gas filtration and purification module connected to the channel. When the battery experiences thermal runaway and releases harmful gases, the channel opens under the control of the safety response control module, guiding the gas in the battery compartment into the filtration and purification module for treatment. The filtration and purification module contains activated carbon adsorbent and catalytic conversion medium, which adsorb toxic gases such as HF and CO and convert them into harmless products, thereby reducing the concentration of harmful gases and the risk of combustion.

[0051] Example 1:

[0052] The system includes: a multi-sensor monitoring module, an AI prediction and evaluation module, a thermal runaway risk assessment module, a safety response control module, and a safety execution module. The multi-sensor monitoring module monitors the operating parameters of the lithium-ion battery, including battery temperature, voltage, current, and gas concentration in the battery compartment environment, acquiring parameter data in real time. The AI ​​prediction and evaluation module performs intelligent analysis and predicts thermal runaway trends based on the operating parameter data, outputting an assessment index of battery thermal runaway risk or a short-term predicted value of battery temperature. The thermal runaway risk assessment module identifies abnormal battery conditions based on operating parameters and prediction and evaluation results, responding to situations where battery temperature and temperature... When the rate of increase exceeds a preset threshold, an abnormal voltage drop trend occurs, or the AI ​​assessment index exceeds the risk threshold, the thermal runaway risk level is determined. The safety response control module is used to trigger corresponding multi-level safety response strategies based on the thermal runaway risk level, and generate control commands to execute predetermined safety response measures, including cooling, power regulation, gas emission treatment, and circuit isolation. The safety execution module includes a cooling device, a gas suppression device, a power-off device, and a fire extinguishing device, which are used to cool the battery, divert and purify harmful gases, cut off circuits, and extinguish fires at different response levels, respectively, to suppress the further expansion of thermal runaway.

[0053] It is worth noting that, specifically, the AI ​​prediction and assessment module uses an artificial intelligence algorithm based on recurrent neural networks to perform time series analysis on multi-source monitoring data of the battery, and outputs the risk probability of lithium battery thermal runaway or temperature trend prediction value. The recurrent neural network may include a long short-term memory (LSTM) network or a lightweight deep neural network, which is used to extract the features of multi-dimensional sensor data changing over time and predict the battery state trend. The system uses the trained long short-term memory neural network (LSTM) model to analyze sensor data such as temperature, voltage, and gas concentration, identify abnormal temperature rising trends or potential signs of thermal runaway in advance, and give a risk score and warning level in combination with a preset threshold strategy.

[0054] Next, specifically, the thermal runaway risk assessment module presets several criteria, including battery temperature threshold, temperature rise rate threshold, voltage drop magnitude threshold, and AI risk score threshold. When monitoring data indicates that the battery temperature or temperature rise rate exceeds the safety threshold, or the voltage change shows an abnormal downward trend, or the risk score output by the AI ​​model exceeds the warning threshold, the system determines that it has entered the corresponding thermal runaway risk condition level.

[0055] Specifically, the multi-level safety response strategy includes at least three response levels: Level 1 warning, Level 2 intervention, and Level 3 emergency protection. In Level 1 warning mode, the system issues an audible and visual alarm and activates the cooling device's normal cooling mode to control the battery's temperature rise. In Level 2 intervention mode, the system reduces the battery's output power to a predetermined safe level, activates enhanced cooling mode, and turns on the gas suppression device to extract and purify the gas inside the battery compartment. In Level 3 emergency protection mode, the system disconnects the battery circuit, isolates the battery from external circuits via a power-off device, and triggers the fire extinguishing device to extinguish any fires on the battery.

[0056] Furthermore, specifically, the cooling device has a regular cooling mode and an emergency cooling mode. When the battery temperature exceeds a first preset threshold, the regular cooling mode is activated to cool the battery; when the temperature continues to rise and exceeds a second preset threshold, it automatically switches to the emergency cooling mode to enhance the cooling intensity. Regular cooling dissipates heat through a fan or liquid cooling circulation, while emergency cooling rapidly suppresses the rise in battery temperature through coolant injection or heat absorption by phase change materials.

[0057] Specifically, the gas suppression device includes a battery compartment depressurization channel and a gas filtration and purification module connected to the channel. When the battery experiences thermal runaway and releases harmful gases, the channel opens under the instruction of the safety response control module, directing the gas in the battery compartment into the filtration and purification module for treatment. The filtration and purification module contains activated carbon adsorbent and catalytic conversion medium, which adsorb toxic gases such as HF and CO and convert them into harmless products, thereby reducing the concentration of harmful gases and the risk of combustion.

[0058] Next, specifically, the safety response control module combines AI prediction results to predict the battery's temperature change trend in the near future; when the prediction results show that the battery temperature will exceed the safety threshold within the set time window, it issues control commands in advance to execute cooling or power limiting measures, thereby intervening in advance before thermal runaway occurs.

[0059] It is worth noting that, specifically, the AI ​​prediction and evaluation module has online learning and model self-optimization capabilities. This module incrementally trains and updates parameters based on battery data collected during operation, enabling the AI ​​model to adapt to battery aging and changes in operating conditions, continuously improving the accuracy and reliability of thermal runaway prediction.

[0060] Finally, specifically, the fire extinguishing device is an aerosol automatic fire extinguishing system, which is installed inside the battery compartment and has multiple spray nozzles arranged towards the battery. When the battery temperature exceeds the danger threshold or an open flame is detected, the safety response control module automatically triggers the aerosol fire extinguishing system to release the extinguishing agent instantly. The extinguishing agent forms uniform fine particles that are suspended around the battery, quickly covering the battery surface and suppressing the spread of the flame.

[0061] It also includes the following steps:

[0062] S1. Acquire multi-source monitoring data of the lithium battery during operation, including parameters such as battery voltage, current, temperature, and gas concentration in the battery compartment;

[0063] S2. Input the monitoring data into the AI ​​prediction and evaluation model for calculation and analysis, evaluate the current state of the battery in real time, and predict the temperature change trend or thermal runaway risk index in a short period of time.

[0064] S3. Compare the monitoring data and AI model prediction results with the preset safety threshold to determine whether the battery shows signs of thermal runaway such as abnormal temperature rise or abnormal voltage drop, and determine the thermal runaway risk level accordingly.

[0065] S4. Based on the determined thermal runaway risk level, invoke the corresponding level of safety response strategy and generate control commands to execute the predetermined safety response operation;

[0066] S5. Execute control commands to implement corresponding safety measures for the battery, including activating the cooling device to lower the temperature, adjusting or cutting off the battery power output, opening the gas suppression channel to discharge and purify harmful gases, and triggering the fire extinguishing device to extinguish the fire, so as to effectively control the potential thermal runaway state of the battery.

[0067] Example 2:

[0068] This invention relates to an AI-based intelligent assessment-based lithium battery thermal runaway prediction and multi-level response system, comprising a multi-sensor monitoring module, an AI prediction and assessment module, a thermal runaway risk assessment module, a safety response control module, and a safety execution module. The multi-sensor monitoring module collects real-time data on battery operation, including temperature, voltage, current, and gas concentration. The AI ​​prediction and assessment module analyzes the battery state and trends using an artificial intelligence model based on the collected data, outputting a real-time thermal runaway risk score or short-term temperature prediction. The thermal runaway risk assessment module identifies abnormal battery conditions based on the prediction results output by the AI ​​model and changes in sensor data; when abnormal trends are detected in parameters such as temperature or voltage, the corresponding thermal runaway risk level is determined. The safety response control module pre-sets multi-level safety response strategies, with different control logics corresponding to different risk levels. Upon receiving a risk level signal from the risk assessment module, it selects and triggers the corresponding level of safety response measures. The safety execution module includes specific execution devices, such as a cooling unit, a ventilation and gas handling unit, a power-off unit, and a fire extinguishing unit, used to perform operations such as cooling, venting, power-off, and fire extinguishing on the battery according to the instructions of the control module. Through the coordinated operation of the above modules, the present invention can take effective detection, control and protection measures at all stages in which thermal runaway of lithium batteries may occur, significantly improving the safety of the battery system.

[0069] The multi-sensor monitoring module includes temperature sensors located inside or on the surface of the battery, voltage sensors connected to the positive and negative terminals of the battery, gas sensors monitoring the battery compartment environment, and current sensors, used to acquire comprehensive battery status information. The temperature sensors can employ thermocouples or thermistors, while the gas sensors detect changes in the concentration of characteristic gases (such as HF and CO) released in the early stages of battery thermal runaway. All sensor signals are connected to a high-speed data acquisition unit to achieve real-time monitoring of battery status parameters.

[0070] The AI ​​prediction and assessment module employs a lightweight artificial neural network model to characterize the dynamic process of battery thermal runaway. For example, a Long Short-Term Memory (LSTM) recurrent neural network with several hidden layers can be used, taking continuously acquired multi-dimensional sensor data as time-series input. Through offline training, the mapping relationship between battery operating conditions and temperature evolution trends is obtained. The trained model is deployed in the battery management system and runs in real time, capable of outputting predicted temperature changes or the probability of thermal runaway occurring within the next tens of seconds. Because the LSTM network can memorize historical information and capture relevant features in the time series, it helps identify abnormal trends that may lead to thermal runaway in the early stages. The introduction of this AI model significantly improves the sensitivity of thermal runaway symptom identification, providing a basis for subsequent decision-making.

[0071] The thermal runaway risk assessment module is equipped with multi-level warning thresholds and judgment rules. For example, a first temperature threshold T1 is set for initial thermal runaway warning, and a second temperature threshold T2 is set for emergency hazard determination (T2 > T1). Temperature rise rate threshold ΔT and voltage drop magnitude threshold ΔV are set to identify abnormal temperature rise rates and voltage drop trends. A risk score threshold P0 output by the AI ​​model is also set for comprehensive evaluation. This module monitors sensor data and AI prediction results in real time. When the temperature or temperature rise rate exceeds the threshold, or an abnormal voltage change occurs, or the AI ​​risk score reaches the threshold P0, the corresponding thermal runaway risk level signal is triggered, notifying the safety response control module to take action.

[0072] The safety response control module maps different risk levels to specific safety response operations according to predefined rules. For each risk level, the control module has a pre-set corresponding control scheme, including the start / stop mode of the cooling device, the adjustment range of battery power output, the start / stop of the gas handling unit, and circuit cut-off and fire extinguishing actions. When the control module receives a risk level signal from the risk assessment module, it immediately sends instructions to each unit of the safety execution module to coordinate the execution of various protective measures according to the established strategy.

[0073] The specific functions of each unit in the safety execution module are as follows: The cooling unit has a variable cooling mode to adjust the heat dissipation intensity as needed. Under normal circumstances, it operates in the conventional cooling mode, maintaining a stable battery temperature through a fan or liquid cooling circulation. When the battery temperature exceeds the threshold T1, the control module instructs the cooling unit to enter an enhanced cooling mode, such as increasing the coolant circulation rate or initiating refrigerant injection to accelerate heat dissipation. If the temperature continues to rise and exceeds a higher threshold T2, further emergency cooling measures (such as releasing the endothermic potential of the phase change material) are taken to quickly suppress the rise in battery temperature. The ventilation and gas handling unit includes a battery compartment depressurization duct and a filtration and purification device. When thermal runaway causes an increase in pressure and harmful gas concentration in the battery compartment, the depressurization duct is opened under control commands to guide the gas in the compartment into the filtration and purification device. The filtration device contains activated carbon and a catalyst, which can adsorb and decompose leaked battery gases such as HF and CO, converting toxic and harmful gases into harmless substances for discharge, preventing the accumulation of flammable gases and reducing environmental harm. The power-off unit is used to quickly disconnect the battery from the external circuit in an emergency, usually achieved by a high-current relay or a burst fuse assembly. When the system determines that it has entered the highest risk level (Level 3 emergency protection), the power-off unit will cut off the battery output circuit within milliseconds to prevent the fault from spreading further. The fire extinguishing unit can use an automatic aerosol fire extinguishing system, with several spray nozzles arranged inside the battery compartment; when an open flame is detected or the battery temperature exceeds the danger threshold, the control module triggers the fire extinguishing unit to instantly release an aerosol extinguishing agent. The extinguishing agent particles rapidly disperse around the battery, forming a uniform suspension covering the battery surface, which can suppress the spread of flames in the initial stage and control the fire in its nascent stage.

[0074] AI Intelligent Assessment and Multi-Level Response Example: In practical applications, this system can dynamically adjust its response strategy based on the prediction results of the AI ​​model, thereby eliminating potential thermal runaway in its early stages. For example, when the temperature sensor of a battery module detects that the temperature is rising at a rate of more than 1°C per second, and the gas sensor detects that the carbon monoxide concentration in the battery compartment is gradually increasing, the AI ​​prediction and assessment module determines that the battery temperature will continue to rise rapidly in the next tens of seconds, and the thermal runaway risk score is approaching the danger threshold. At this time, the risk assessment module directly upgrades the system risk level to Level 2 intervention. The safety response control module then issues instructions: on the one hand, it limits the battery discharge current to a safe range to reduce internal heat generation; on the other hand, it immediately activates the enhanced mode of the cooling unit to accelerate heat dissipation and opens the ventilation ducts to extract and purify the gas in the battery compartment. After the above intervention, the rise in battery temperature slows down. However, if the temperature is still not effectively controlled and continues to approach the T2 threshold, the system will further upgrade to Level 3 emergency protection: the instantaneous trigger power-off unit cuts off the battery output to prevent heat and current from continuing to accumulate, and at the same time, the fire extinguishing unit is activated to spray aerosol fire extinguishing agent around the battery to suppress any possible open flames. Through this intelligent hierarchical response mechanism, the entire process is effectively controlled before thermal runaway actually occurs, eliminating potential accidents in their infancy and significantly improving the intrinsic safety level of lithium battery systems.

[0075] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A lithium battery thermal runaway prediction and multi-level response system based on AI intelligent evaluation, characterized in that, The system comprises a multi-sensor monitoring module, an AI prediction and evaluation module, a thermal runaway risk assessment module, a safety response control module, and a safety execution module. The multi-sensor monitoring module is used to monitor the operating parameters of the lithium-ion battery, including battery temperature, voltage, current, and gas concentration in the battery compartment environment, and to obtain real-time parameter data. The AI prediction and evaluation module is used to perform intelligent analysis and thermal runaway trend prediction based on the operating parameter data, outputting an evaluation index of battery thermal runaway risk or a short-term prediction value of battery temperature. The thermal runaway risk assessment module is used to identify abnormal conditions of the battery based on the operating parameters and the prediction and evaluation results. When the battery temperature and temperature rise rate exceed the preset threshold, an abnormal voltage drop trend occurs, or the AI evaluation index exceeds the risk threshold, the thermal runaway risk level is determined. The safety response control module is used to trigger corresponding multi-level safety response strategies according to the thermal runaway risk level, generate control instructions to execute predetermined safety response measures, including cooling, power adjustment, gas discharge treatment, and circuit isolation. The safety execution module includes a cooling device, a gas suppression device, a power-off device, and a fire extinguishing device, which are used to cool the battery, purify harmful gas, cut off the circuit, and extinguish the fire at different response levels to suppress the further expansion of thermal runaway.

2. The lithium battery thermal runaway prediction and multi-stage response system based on AI intelligent evaluation according to claim 1, characterized in that, The AI prediction and evaluation module uses a recurrent neural network-based algorithm to perform time series analysis on the multi-source monitoring data, outputting a risk probability or temperature trend prediction value of lithium battery thermal runaway. The recurrent neural network includes a long short-term memory (LSTM) network or a lightweight deep neural network, which is used to train and learn the continuously monitored data to extract time series features. 3.The lithium battery thermal runaway prediction and multi-stage response system based on AI intelligent evaluation of claim 2, wherein The thermal runaway risk assessment module has multiple preset determination conditions, including battery temperature threshold, temperature rise rate threshold, voltage drop amplitude threshold, and AI risk score threshold. When the battery temperature or temperature rise rate exceeds the safety threshold, or the battery voltage shows an abnormal downward trend, or the AI model output risk score exceeds the risk score threshold, the corresponding thermal runaway risk level is determined and a risk alarm signal is output.

4. The lithium battery thermal runaway prediction and multi-stage response system based on AI intelligent evaluation according to claim 3, characterized in that, The multi-level safety response strategy includes at least three response levels: first-level warning, second-level intervention, and third-level emergency protection. In the first-level warning state, the system issues an audible and visual alarm and starts the regular cooling mode of the cooling device to control the battery temperature rise trend. In the second-level intervention state, the system reduces the battery power output to a predetermined safety level, starts the enhanced cooling mode, and turns on the gas suppression device to exhaust and purify the gas in the battery compartment. In the third-level emergency protection state, the system cuts off the connection between the battery and the external circuit and triggers the fire extinguishing device to extinguish the fire in the battery.

5. The lithium battery thermal runaway prediction and multi-stage response system based on AI intelligent evaluation according to claim 4, characterized in that, The AI prediction and evaluation module has online learning and model self-adaptive optimization functions. By incrementally training the battery data collected during operation to update the model parameters, the artificial intelligence model can continuously improve the accuracy of thermal runaway risk prediction as the battery ages and the working conditions change.

6. The lithium battery thermal runaway prediction and multi-stage response system based on AI intelligent evaluation according to claim 5, characterized in that, The cooling device has a normal cooling mode and an emergency cooling mode; the normal cooling mode is started to reduce the temperature when the battery temperature exceeds a first preset threshold, and the emergency cooling mode is automatically switched to when the temperature continues to rise and exceeds a second preset threshold to enhance the cooling intensity, wherein the normal cooling is conducted by heat dissipation through a fan or liquid cooling circulation, and the emergency cooling is conducted by rapid inhibition of the battery temperature rise through coolant injection or phase change material heat absorption.

7. The lithium battery thermal runaway prediction and multi-stage response system based on AI intelligent evaluation according to claim 6, characterized in that, The gas suppression device comprises a battery cabin pressure relief flow guide channel and a gas filtration and purification module connected with the flow guide channel; when the battery releases harmful gas due to thermal runaway, the flow guide channel is opened under the control of a safety response control module to guide the gas in the battery cabin into the filtration and purification module for treatment; the filtration and purification module contains activated carbon adsorbent and catalytic conversion medium for adsorbing toxic gases such as HF and CO and converting them into harmless products, thereby reducing the concentration of harmful gas and the risk of combustion.

Images