Multi-stage thermal runaway suppression method and device based on integrated power battery

Abstract

The invention discloses a multistage thermal runaway suppression method and device based on an integrated power battery. The method comprises the steps that power battery data information implemented by a vehicle is acquired; inputting the power battery data into a preset thermal runaway prediction model based on an LSTM neural network; outputting a real-time risk probability through a thermal runaway prediction model based on the LSTM neural network; according to the thermal runaway prediction model based on the LSTM neural network, on the basis of the real-time risk probability, according to current preset threshold value judgment, the levels of thermal runaway suppression are output, and the suppression levels include first-level suppression, second-level suppression and third-level suppression. The thermal runaway suppression system has the beneficial effects that rapid cooling and fire extinguishing are achieved through first-stage suppression, thermal diffusion is prevented through second-stage suppression physical blocking, safe temperature is maintained through third-stage suppression, reburning is prevented, progressive layer by layer is achieved, the success rate, reliability and comprehensiveness of thermal runaway suppression are remarkably improved, and the risk of fire and explosion of a battery pack is greatly reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of vehicle thermal runaway, and in particular to a multi-stage thermal runaway suppression method and device based on integrated power battery. BACKGROUND

[0002] With the rapid development of new energy vehicle industry, the capacity and energy density of power battery are continuously improved. However, high energy density battery is prone to cause chain reaction when thermal runaway occurs, leading to vehicle fire and even explosion, which seriously threatens the safety of vehicle and passengers. Therefore, the industry generally pays attention to the early warning and suppression technology of battery thermal runaway.

[0003] In the prior art, a thermal runaway early warning and suppression system is usually adopted, which includes a battery management unit (BMS), a thermal runaway early warning unit and an external thermal runaway suppression unit. This kind of scheme monitors the temperature, voltage and flammable gas concentration of the battery through sensors, and starts the suppression system when an abnormality is found. The common measure is to use a sprayer to inject fire extinguishing agent into the battery module, thereby reducing the battery temperature and suppressing the spread of fire. This kind of technology can slow down the spread of thermal runaway to a certain extent and reduce the risk of accidents.

[0004] However, the above prior art still has the following shortcomings:

[0005] The current scheme relies on single fire extinguishing agent injection, which can quickly reduce the temperature, but has obvious differences in different stages (initial stage, middle stage and residual heat stable stage) of thermal runaway, and it is difficult to achieve full-cycle and all-round protection. Especially in the later stage of thermal runaway, the fire extinguishing agent alone is not enough to prevent heat conduction between the battery cells, and there is a risk of secondary combustion and "domino effect".

[0006] In the process of thermal runaway, a large amount of heat is released in the form of waste energy. The existing technology only focuses on fire extinguishing and suppression, and does not consider recycling and reusing the residual heat, resulting in energy waste. At the same time, the alarm, heat dissipation and electric control action in the process of thermal runaway disposal need to consume a large amount of electric energy, and once the main battery system is damaged, the key components (such as braking and steering) may cause greater safety hazards due to lack of power support. SUMMARY

[0007] The purpose of the present application is to solve the technical problems in the prior art and provide a multi-stage thermal runaway suppression method, device, equipment and medium for integrated power battery of 48V system vehicle.

[0008] The present application relates to a multi-stage thermal runaway suppression system for integrated power battery of 48V system vehicle, an electronic device, a computer readable storage medium and a computer program product, to solve the technical defects in the prior art.

[0009] Technical solutions:

[0010] In a first aspect, the application provides a multi-stage thermal runaway suppression method based on an integrated power battery, comprising:

[0011] Obtaining power battery data information implemented by a vehicle;

[0012] Inputting the power battery data into a preset thermal runaway prediction model based on an LSTM neural network;

[0013] Outputting a real-time risk probability through the thermal runaway prediction model based on the LSTM neural network;

[0014] Wherein, the thermal runaway prediction model based on the LSTM neural network outputs a suppression level of thermal runaway according to a current preset threshold based on the real-time risk probability, and the suppression levels are first-level suppression, second-level suppression and third-level suppression;

[0015] Different suppression strategies are implemented through corresponding suppression levels.

[0016] Preferably, the power battery data information includes temperature and voltage data.

[0017] Preferably, the thermal runaway prediction model based on the LSTM neural network is trained by an early-acquired power battery thermal runaway data set, and outputs a 30-second future risk probability and a corresponding suppression level.

[0018] Preferably, the thermal runaway prediction model based on the LSTM neural network comprises:

[0019] Obtaining temperature and voltage data, and collecting the temperature change rate with time and the voltage drop speed with time through an edge computing node;

[0020] The time sequence feature formula is calculated as follows:

[0021] ;

[0022] The risk probability output formula is calculated as follows:

[0023] ;

[0024] Wherein, ΔT / Δt is the temperature change rate with time, ΔV / Δt is the voltage drop speed with time, σ is the Sigmoid activation function, P∈[0,1] is ensured, t is the current calculation time, k is the time window length, W and b are model training parameters, and n is the time window length.

[0025] Preferably, the thermal runaway prediction model based on the LSTM neural network outputs the level of thermal runaway suppression according to the current preset threshold value based on the real-time risk probability, and the suppression levels are first-level suppression, second-level suppression and third-level suppression, including:

[0026] According to the numerical range of the risk probability, a multi-level suppression mechanism is triggered in turn:

[0027] When P>k2, start the first-level suppression, and spray the aerosol fire extinguishing agent;

[0028] When k1≤P≤k2, start the second-level suppression, and form a self-repairing isolation layer;

[0029] When P<k1, and after starting the first-level suppression or the second-level suppression, enable the third-level suppression, and perform dynamic temperature control and waste heat recovery;

[0030] Wherein, k1 is the first threshold value, and k2 is the second threshold value.

[0031] Preferably, different suppression strategies are carried out through the corresponding suppression levels, including: when the first-level suppression is started, the aerosol fire extinguishing agent is triggered to spray;

[0032] Which includes:

[0033] Send a pulse to the piezoelectric valve that triggers the aerosol fire extinguishing agent to spray the aerosol fire extinguishing agent on the power battery, and calculate the spraying time, the formula is as follows:

[0034] ;

[0035] Wherein, T instant is the instantaneous temperature, T safe is the safety threshold temperature, k is the temperature coefficient, and η is the fire extinguishing agent efficiency coefficient.

[0036] Preferably, different suppression strategies are carried out through the corresponding suppression levels, including: when the second-level suppression is started, a self-repairing isolation layer is formed;

[0037] Which includes:

[0038] Control the paraffin and expanded graphite composite material spraying device to spray the power battery;

[0039] Wherein, the spraying material contains photosensitive resin microcapsule repair agent.

[0040] Preferably, different suppression strategies are carried out through the corresponding suppression levels, including: when the third-level suppression is started, dynamic temperature control and waste heat recovery are performed;

[0041] Which includes:

[0042] The PID algorithm is triggered to dynamically adjust the driving voltage of the thermoelectric material on the gradient distribution of the starting power battery, so as to maintain the temperature difference ΔT of the thermoelectric material sheet at 50-80℃ in the high-efficiency interval, and the driving voltage formula is as follows:

[0043]

[0044]

[0045] The thermoelectric conversion efficiency is calculated in real time, and when η 热电 <15%, the Vte is automatically adjusted to restore to the high-efficiency interval;

[0046] The conversion efficiency formula is as follows:

[0047]

[0048] 热电 η q is the heat energy passing through the thermoelectric material per unit time,

[0049] The recovered electric energy is distributed by the energy management module, and is preferentially supplied to the PTC heater for dynamically adjusting the power battery temperature, so as to maintain the battery module in a safe temperature range, and is secondarily supplied to the super capacitor and the steering system.

[0050] In a second aspect, an embodiment of the present application provides a multi-stage thermal runaway suppression system based on an integrated power battery, comprising:

[0051] An acquisition unit acquires power battery data information of a vehicle;

[0052] A detection unit inputs the power battery data into a preset thermal runaway prediction model based on an LSTM neural network;

[0053] The thermal runaway prediction model based on the LSTM neural network outputs a real-time risk probability;

[0054] The thermal runaway prediction model based on the LSTM neural network outputs a real-time risk probability;

[0055] ​​​​​​​​​​​The execution unit is used for different suppression strategies by corresponding suppression levels.

[0056] In a third aspect, an electronic device is provided, including a processor and a memory.

[0057] In a fourth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program.

[0058] In a fifth aspect, a computer program product is provided, and when the computer program product is run on an electronic device, the electronic device executes the method of any possible design of the above-mentioned aspects.

[0059] Beneficial effects: The application breaks through the past single fire extinguishing technical route, and innovatively designs a multi-level synergistic suppression mechanism of a first-level directional aerosol fire extinguishing, a second-level self-repairing isolation layer and a third-level dynamic temperature control isolation belt. The mechanism is based on an advanced LSTM neural network prediction model, and different levels are accurately triggered through real-time risk probability (P value) to realize the coverage protection of the whole process of the thermal runaway "burst period", "spreading period" and "thermal stable period". The first-level suppression rapidly reduces the temperature and extinguishes the fire, the second-level suppression prevents the thermal diffusion through physical blocking, and the third-level suppression maintains the safe temperature and prevents the re-ignition, which layer by layer progresses, significantly improves the success rate, reliability and comprehensiveness of the thermal runaway suppression, and greatly reduces the risk of battery pack fire explosion.

[0060] The application first deeply integrates the thermoelectric conversion technology (Bi2Te3) in the thermal runaway suppression system. In the third-level suppression stage, the system can convert the dangerous residual heat into electric energy, and feedback to the 48V vehicle system through a bidirectional DC-DC converter, so as to realize the efficient recycling of energy. This not only greatly reduces the net energy consumption in the suppression process, improves the economy and environmental protection of the whole system, and more importantly, the recycled electric energy can provide emergency power supply support for the key safety systems of the vehicle according to the priority. Even in the extreme case of thermal runaway leading to the failure of the main power supply, the braking and steering functions of the vehicle can still be guaranteed, which provides valuable time for the passengers to evacuate and enhances the safety redundancy of the whole vehicle.

[0061] The improved LSTM prediction model adopted by the present application can predict the thermal runaway risk probability within 30 seconds in the future based on historical data by introducing time sequence characteristics such as voltage drop rate (AV / At), and the response delay is less than 50 ms. The prediction ability based on artificial intelligence enables the system to make a judgment and start the suppression measure at an extremely early stage or at the moment of thermal runaway, greatly improving the initiative and effectiveness of protection. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 A method framework schematic diagram is provided for the present application;

[0063] Figure 2 A system schematic diagram is provided for the present application;

[0064] Figure 3 It is a device structure block diagram provided by an embodiment of the present application.

[0065] Figure 4 It is an electronic device structure block diagram provided by an embodiment of the present application. DETAILED DESCRIPTION

[0066] In order to make the technical solutions of the present application clearer, the following will further specifically describe the present application with specific embodiments combined with the drawings.

[0067] Embodiment 1

[0068] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely in the following combined with the drawings of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application. Unless otherwise defined, the technical terms or scientific terms used herein should be understood as the usual meanings understood by those skilled in the art in the field of the present application. The words such as "include" and similar words used herein mean that the elements or objects before the words cover the elements or objects listed after the words and their equivalents, without excluding other elements or objects.

[0069] In view of the problems existing in the prior art, such as Figure 1 , a multi-stage thermal runaway suppression method based on integrated power battery is proposed, which includes

[0070] Step S1, acquiring power battery data information of the vehicle, in some specific embodiments, the power battery data information includes temperature, voltage data, real-time collection of temperature, voltage, etc. Operating parameters of the battery are used as basic data for thermal runaway prediction;

[0071] Step S2, input the power battery data into the preset thermal runaway prediction model based on LSTM neural network; in some specific embodiments, the thermal runaway prediction model based on LSTM neural network is trained by the thermal runaway data set obtained in advance, and outputs the risk probability and the corresponding inhibition level in the next 30 seconds. The improved LSTM model is used to process time series data (temperature gradient, voltage drop rate, etc.), and output the thermal runaway risk probability in a period of time (such as 30 seconds) in the future.

[0072] The real-time risk probability is output by the thermal runaway prediction model based on LSTM neural network;

[0073] The thermal runaway inhibition level is output by the thermal runaway prediction model based on LSTM neural network based on the real-time risk probability according to the current preset threshold value, and the inhibition levels are first inhibition, second inhibition and third inhibition; the model not only outputs the risk probability, but also directly gives the recommended inhibition level (first, second and third) according to the preset threshold value. The additional judgment link is omitted, the prediction and decision are integrated, and the response speed is accelerated.

[0074] Step S3, different inhibition strategies are carried out according to the corresponding inhibition level.

[0075] In some specific embodiments, the thermal runaway prediction model based on LSTM neural network comprises:

[0076] Temperature and voltage data are obtained, and the rate of change of temperature with time and the rate of decline of voltage with time are collected by the edge computing node;

[0077] The time series feature formula is as follows:

[0078] ;

[0079] The risk probability output formula is as follows:

[0080] ;

[0081] Wherein, ΔT / Δt is the rate of change of temperature with time, the core of thermal runaway is chain exothermic reaction, and the temperature will rise sharply. ΔT / Δt can sensitively capture this abnormal and accelerated temperature rising trend, and give an early warning much earlier than the absolute temperature value. The voltage drop rate ΔV / Δt is the rate of decline of voltage with time, and when the battery internal short circuit, diaphragm collapse and other faults occur, the output voltage will drop rapidly. This feature is a strong signal of irreversible damage inside the battery, and σ is the Sigmoid activation function, wherein, , to ensure that P∈[0,1], t is the current calculation time, k is the time window length, W and b are model training parameters, and n is the time window length.

[0082] Specifically, the comprehensive utilization of two types of key features (temperature gradient, voltage drop rate) is more accurate than single temperature monitoring, LSTM can remember historical trends, and is suitable for battery thermal runaway which is a gradual process. Compared with traditional threshold judgment, the continuous risk probability is given, which is convenient for setting a hierarchical suppression strategy.

[0083] Thermal runaway is a dynamic development process, and the current state is highly dependent on the historical state of the previous period. LSTM (Long Short-Term Memory Network) is a typical model for processing such time series data, which is good at learning long-term dependencies from historical data and predicting future trends;

[0084] The model does not only look at the change rate at the current time point, but analyzes the continuous data in a time window from t-n time to the current t time. This ensures that the model can make judgments based on trends over a period of time, rather than being disturbed by instantaneous noise, greatly improving the reliability of the prediction.

[0085] The output is compressed to the [0, 1] interval through the Sigmoid function, forming an intuitive risk probability value P.

[0086] For example, P=0 means absolute safety.

[0087] P=1 means that thermal runaway will definitely occur.

[0088] P=0.85 means there is an 85% chance of thermal runaway occurring in the future.

[0089] This continuous probability output contains much more information than a simple "yes / no" binary alarm. It provides accurate and quantifiable decision-making basis for subsequent multi-level triggering strategies (for example, P>0.8 triggers level one, 0.6≤P≤0.8 triggers level two).

[0090] The data is collected and processed by the edge computing node. Processing near the data source avoids network delays caused by uploading data to the central controller, meeting the stringent requirements of millisecond-level response for thermal runaway suppression. Even if the vehicle network fails, the edge node can work independently to perform local calculation and decision-making, ensuring system redundancy and safety. Dispersing complex model calculation tasks to the edge avoids the central controller's algorithmic bottleneck.

[0091] In some specific embodiments, the thermal runaway prediction model based on the LSTM neural network outputs the level of thermal runaway suppression according to the current preset threshold based on the real-time risk probability, and the suppression levels are first-level suppression, second-level suppression, and third-level suppression, including:

[0092] According to the numerical range of risk probability, a multi-level suppression mechanism is triggered in sequence:

[0093] When P>k2, the first level of suppression is started, and aerosol fire extinguishing agent is sprayed;

[0094] When k1≤P≤k2, the second level of suppression is started, and a self-repairing isolation layer is formed;

[0095] When P<k1, and after the first or second level of suppression is started, the third level of suppression is enabled, and dynamic temperature control and residual heat recovery are performed;

[0096] Where k1 is the first threshold value, and k2 is the second threshold value.

[0097] Specifically, in some examples, k1 can be 0.6 and k2 can be 0.8;

[0098] P is a value between 0 and 1, calculated by the LSTM model, representing the likelihood of thermal runaway occurring in the near future. It is a predictive and forward-looking indicator that signals earlier than any current physical quantity monitored, gaining valuable time for intervention.

[0099] Multi-level threshold judgment: k1 and k2;

[0100] K2 (second threshold value, e.g. 0.8): This is a high-risk threshold. Once exceeded, it means that thermal runaway has broken out or is about to break out, requiring the most intense and fastest fire extinguishing means (first level of suppression).

[0101] K1 (first threshold value, e.g. 0.6): This is a medium-risk threshold. Between k1 and k2, it means that the risk is accumulating and spreading, and physical isolation measures need to be taken to prevent the situation from getting worse (second level of suppression).

[0102] P<k1 (e.g. P<0.6): This means that the thermal runaway risk has been initially controlled and entered a thermal stable period. The task at this time is to manage residual heat and recover energy (third level of suppression).

[0103] In some specific embodiments, different suppression strategies are carried out through corresponding suppression levels, including: when the first level of suppression is started, aerosol fire extinguishing agent is sprayed;

[0104] This includes:

[0105] Send a pulse to the piezoelectric valve that triggers the aerosol fire extinguishing agent, spray aerosol fire extinguishing agent on the power battery, and calculate the spraying duration, formula as follows:

[0106] ;

[0107] Where T instant is the instantaneous temperature, Tsafe T is the safety threshold temperature, k is the temperature coefficient, and η is the extinguishing agent efficiency coefficient.

[0108] Specifically, sending a pulse to the piezoelectric valve is an execution action. The control logic module sends a high-frequency PWM pulse signal (such as the previously mentioned 20 kHz) to drive the piezoelectric valve. The piezoelectric valve has the advantage of extremely fast response (<1 ms), enabling instantaneous opening and precise control of the spray.

[0109] The dynamic calculation of the spray duration t is the decision logic. The duration of the spray is not fixed, but is intelligently calculated based on the severity of the fire (the difference between the instantaneous temperature and the safety temperature).

[0110] T instant (Temporary temperature): When thermal runaway occurs, the highest temperature on the surface of the battery cell is collected in real time by the sensor. This represents the severity of the current thermal runaway.

[0111] T safe (Safety threshold temperature): A pre-set safety target temperature, which is the critical temperature at which the battery can stop the exothermic reaction and return to normal working condition. For example, it may be 60°C or 70°C.

[0112] T instant -T safe (Temperature difference): This is the heat that the system needs to "eliminate". The larger the difference, the more serious the fire, the more extinguishing agent is needed, and the longer the spray time naturally is.

[0113] k (temperature coefficient): An empirical constant (specifically, it can be set to k = 0.35°C / ms). Its physical meaning can be understood as the temperature reduction per second of the extinguishing agent. The larger the k value, the higher the extinguishing agent cooling efficiency, and the shorter the required spray time t.

[0114] η (extinguishing agent efficiency coefficient) This coefficient reflects the overall efficiency of the extinguishing agent formula, spray pressure, coverage uniformity, etc.

[0115] Primary suppression is rapidly initiated within 0-3 seconds after thermal runaway occurs, and an aerosol extinguishing agent is sprayed through a piezoelectric valve (flow rate 1.2 L / s, spray distance 40 cm, response time <1 ms).

[0116] The extinguishing agent formula includes perfluorohexanone (85%), nano-aluminum oxide (10%), and corrosion inhibitor (5%), which can quickly cover the surface of the battery module, reduce the temperature, and prevent battery corrosion. The spray path adopts a variable angle design (30°-60° adaptive), covering an area ≥150% of the module surface area, ensuring that the extinguishing agent can be evenly distributed to quickly suppress thermal runaway.

[0117] In some specific embodiments, different suppression strategies are carried out through corresponding suppression levels, including: starting secondary suppression, forming a self-repairing isolation layer;

[0118] Among them:

[0119] The paraffin and expanded graphite composite material jetting device is used to jet the power battery;

[0120] Among them, the jetting material contains photosensitive resin microcapsule repair agent.

[0121] Specifically, paraffin as a phase change material, it will melt and absorb heat at a certain temperature (such as 70℃), which can effectively absorb a large amount of heat generated by battery thermal runaway, and play a role in auxiliary cooling.

[0122] Expanded graphite is the main body of the isolation layer. When it encounters high temperature, expanded graphite will expand rapidly by hundreds of times (expansion coefficient ≥250), becoming a fluffy, worm-like graphite flake accumulation body. This accumulation body:

[0123] Has excellent thermal insulation, which can effectively prevent high temperature from conducting to adjacent battery modules.

[0124] Has good coverage and inertia, which can isolate air (oxygen) and suppress flame combustion.

[0125] Has good flexibility, which can be wrapped around the surface of the battery and adapt to irregular shapes.

[0126] Microcapsule technology wraps an uncured photosensitive resin in extremely small capsules, which are then uniformly mixed into the above-mentioned composite material. When the isolation layer cracks or is damaged due to mechanical vibration, impact or other reasons, these microcapsules will break. The pre-set ultraviolet light (or other UV source) in the vehicle battery box will irradiate the damaged area. The photosensitive resin flowing out of the broken capsules will quickly solidify under ultraviolet irradiation, thereby re-adhering the cracks and achieving automatic repair. This ensures the integrity and durability of the isolation layer, avoiding the risk of thermal runaway spreading again due to damage to the isolation layer, greatly improving the reliability of the system.

[0127] The isolation layer contains microcapsule repair agent (photosensitive resin), which triggers self-repair through ultraviolet light when damaged, with a response time of <10s. The thickness control accuracy of the isolation layer is ±0.1mm, and the porosity is <3%, which can effectively prevent heat conduction and further suppress the spread of thermal runaway.

[0128] In some specific embodiments, different suppression strategies are carried out through corresponding suppression levels, including: starting secondary suppression, forming a self-repairing isolation layer;

[0129] Among them:

[0130] The PID algorithm is triggered to dynamically adjust the driving voltage of the thermoelectric material (Bi2Te3) on the gradient distribution of the starting power battery to maintain the temperature difference ΔT between the two ends of the thermoelectric material sheet in the high-efficiency interval of 50-80℃, and the driving voltage formula is as follows:

[0131] ;

[0132] wherein, is the driving voltage of the thermoelectric material, is the proportional coefficient, is the integral coefficient, is the current highest temperature of the battery, is the target temperature, and t is time;

[0133] The thermoelectric conversion efficiency is calculated in real time, and when η 热电 <15%, Vte is automatically adjusted to restore to the high-efficiency interval;

[0134] The conversion efficiency formula is as follows:

[0135] ;

[0136] η 热电 is the conversion efficiency of the waste heat recovery electric energy, is the current of the thermoelectric material, is the terminal voltage of the thermoelectric material, is the contact area of the thermoelectric material and the power battery, and q is the heat energy passing through the thermoelectric material per unit time,

[0137] The recovered electric energy is distributed by the energy management module, and is preferentially supplied to the PTC heater for dynamically adjusting the power battery temperature, maintaining the battery module within the safe temperature range, and secondarily supplied to the super capacitor and the steering system.

[0138] Specifically, when the risk probability P<0.6, the system enters the thermal stable period, the PID algorithm adjusts the driving voltage V te , and maintains the temperature difference ΔT between the two ends of the thermoelectric sheet in the high-efficiency interval of 50-80℃. Real-time dynamic adjustment is avoided to avoid excessive or excessive temperature. The thermoelectric conversion efficiency is calculated when η 热电 <15%, V te is automatically adjusted to restore to the high-efficiency state. The recovered electric energy is distributed by the energy management module: preferentially supplied to the PTC heater (for dynamically adjusting the battery temperature, weight 60%). Secondly, supply the super capacitor (10%) and the steering system (weight 30%) to guarantee the key safety function.

[0139] In some specific embodiments, in combination with Figure 2Also proposed is a multi-stage thermal runaway suppression system based on an integrated power battery, comprising an information acquisition module for acquiring power battery data, a power battery, a first-stage suppression module, a second-stage suppression module, a third-stage suppression module, an energy management module, a 48V system and emergency power supply module, and a control logic module for implementing the method of the above embodiment.

[0140] In some specific embodiments, the energy management module manages the waste heat electric energy recovered by the thermoelectric material (Bi2Te3). The voltage is boosted or bucked by a bidirectional DC-DC converter, ensuring efficient energy feedback to the 48V system; the energy storage unit (supercapacitor or auxiliary battery) is used for energy storage and emergency discharge. During the thermal runaway disposal process, the recovered electric energy can be stably managed, and the output can be ensured to be smooth. The waste heat generated by the battery thermal runaway is converted into electric energy and efficiently utilized. The recovered electric energy is dynamically allocated, and the temperature control and key safety modules are prioritized. Energy waste is reduced, and safety and economy are considered.

[0141] In some specific embodiments, the 48V system and emergency power supply module provides emergency power support for other key systems of the vehicle. The power battery system is the core power supply platform, which is connected to key systems such as braking and steering. During the thermal runaway suppression process, even if the main power battery is damaged, the key systems can still operate. In an emergency, the core systems such as braking and steering are ensured to be powered on. The vehicle is prevented from losing control due to battery failure. The recovered electric energy is preferentially supplied to the key systems, and the emergency capability of the vehicle is improved.

[0142] In another embodiment of the present application, in combination with 3, the embodiment of the present application discloses a multi-stage thermal runaway suppression system based on an integrated power battery, comprising:

[0143] The acquisition unit 301 acquires power battery data information implemented by the vehicle;

[0144] The detection unit 302 is configured to input the power battery data into a thermal runaway prediction model based on a preset LSTM neural network;

[0145] The thermal runaway prediction model based on the LSTM neural network outputs a real-time risk probability;

[0146] The thermal runaway prediction model based on the LSTM neural network outputs a real-time risk probability;

[0147] The execution unit 303 is configured to implement different suppression strategies according to the corresponding suppression levels.

[0148] All related contents of each step involved in the above method embodiment can be cited to the function description of the corresponding function module, which will not be repeated here.

[0149] In some embodiments of the present application, an electronic device 400 is disclosed, as shown in Figure 4 application programs (not shown) and one or more computer programs 404, which can be connected through one or more communication buses 405. The one or more computer programs 404 are stored in the above-mentioned memory 402 and configured to be executed by the one or more processors 401, and the one or more computer programs 404 include instructions that can be used to perform various steps in the embodiments of the present application and corresponding embodiments. Figure 1

[0150] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of functional modules is taken as an example, and in actual application, the above-mentioned functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0151] The functional units in each of the embodiments of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0152] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor execute all or part of the steps of the methods described in the embodiments of the present application. The foregoing storage medium includes: a flash memory, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0153] ​The above merely describes specific implementation of the embodiments of the present application, but the protection scope of the embodiments of the present application is not limited thereto, any change or replacement within the technical scope disclosed by the embodiments of the present application should be covered in the protection scope of the embodiments of the present application. Therefore, the protection scope of the embodiments of the present application should be subject to the protection scope of the claims.

Claims

1. A multi-stage thermal runaway suppression method based on integrated power battery, characterized in that, The method comprises the following steps: acquiring power battery data information implemented by a vehicle; inputting the power battery data into a preset thermal runaway prediction model based on an LSTM neural network; outputting a real-time risk probability through the thermal runaway prediction model based on the LSTM neural network; wherein the thermal runaway prediction model based on the LSTM neural network outputs a level of thermal runaway suppression according to a current preset threshold value based on the real-time risk probability, and the level of suppression is first-level suppression, second-level suppression and third-level suppression; and different suppression strategies are implemented through the corresponding levels of suppression.

2. The method of claim 1, wherein, The power battery data information includes temperature and voltage data; The thermal runaway prediction model based on the LSTM neural network is trained by an early-acquired power battery thermal runaway data set, and outputs a 30-second future risk probability and a corresponding suppression level.

3. The method of claim 2, wherein, The thermal runaway prediction model based on the LSTM neural network comprises the following steps: acquiring temperature and voltage data, and collecting the temperature change rate with time and the voltage drop speed with time through an edge computing node; the time sequence feature formula is calculated as follows: ； the risk probability output formula is calculated as follows: ； wherein ΔT / Δt is the temperature change rate with time, ΔV / Δt is the voltage drop speed with time, σ is a Sigmoid activation function, P ∈ [0, 1] is ensured, t is the current calculation time, k is the time window length, W and b are model training parameters, and n is the time window length.

4. The method of claim 1, wherein, The thermal runaway prediction model based on the LSTM neural network outputs a level of thermal runaway suppression according to a current preset threshold value based on the real-time risk probability, and the level of suppression is first-level suppression, second-level suppression and third-level suppression, including the following steps: according to the numerical range of the risk probability, a multi-level suppression mechanism is triggered in sequence: when P > k2, first-level suppression is started, and an aerosol extinguishing agent is sprayed; when k1 ≤ P ≤ k2, second-level suppression is started, and a self-repairing isolation layer is formed; when P < k1, and after first-level suppression or second-level suppression is started, third-level suppression is enabled, and dynamic temperature control and waste heat recovery are performed; wherein k1 is a first threshold value, and k2 is a second threshold value.

5. The method of claim 4, wherein, Different suppression strategies are implemented through the corresponding levels of suppression, including the following steps: when first-level suppression is started, an aerosol extinguishing agent is sprayed; which comprises the following steps: ； where T instant is the instantaneous temperature, T safe is the safety threshold temperature, k is the temperature coefficient, and η is the extinguishing agent efficiency coefficient. a pulse is sent to a piezoelectric valve that triggers the spraying of the aerosol extinguishing agent, the aerosol extinguishing agent is sprayed on the power battery, and the spraying duration is calculated, and the formula is as follows:

6. The method of claim 5, wherein, wherein the extinguishing agent formula includes perfluorohexanone, nano-aluminum oxide and corrosion inhibitor. Different suppression strategies are implemented through the corresponding levels of suppression, including the following steps: second-level suppression is started, and a self-repairing isolation layer is formed; which comprises the following steps:

7. The method of claim 5, wherein, a paraffin and expanded graphite composite material spraying device sprays on the power battery; wherein the spraying material contains photosensitive resin microcapsule repair agent. Different suppression strategies are implemented through the corresponding levels of suppression, including the following steps: ； wherein, is the thermoelectric material drive voltage, is a proportional coefficient, is an integral coefficient, is the current maximum temperature of the battery, is the target temperature, t is time; Real-time calculation of the thermoelectric conversion efficiency, when η 热电 <15%, automatic adjustment of Vte to restore to the high efficiency interval; third-level suppression is started, and dynamic temperature control and waste heat recovery are performed; which comprises the following steps: a PID algorithm is triggered to dynamically adjust the driving voltage of the thermoelectric material on the power battery to maintain the temperature difference ΔT between the two ends of the thermoelectric material sheet in the high-efficiency interval of 50-80 ℃, and the driving voltage formula is as follows: the conversion efficiency formula is as follows: ； η 热电 η is the conversion efficiency of the waste heat recovery electric energy, I is the current generated by the thermoelectric material, V is the terminal voltage of the thermoelectric material, A is the contact area of the thermoelectric material and the power battery, q is the heat energy passing through the thermoelectric material per unit time, and ΔT is the temperature difference on both sides of the thermoelectric material. The recovered electric energy is distributed by the energy management module, and is preferentially supplied to the PTC heater for dynamically adjusting the temperature of the power battery, so as to maintain the battery module in a safe temperature range, and is secondarily supplied to the super capacitor and the steering system.

8. A multi-stage thermal runaway mitigation system based on integrated power battery, characterized in that, The method comprises the following steps: an acquisition unit acquires power battery data information implemented by a vehicle; a detection unit inputs the power battery data into a preset thermal runaway prediction model based on an LSTM neural network; an output real-time risk probability is output by the thermal runaway prediction model based on the LSTM neural network; wherein the thermal runaway prediction model based on the LSTM neural network outputs a level of thermal runaway suppression according to a current preset threshold value based on the real-time risk probability, and the suppression level is respectively a first suppression, a second suppression and a third suppression; an execution unit implements different suppression strategies according to the corresponding suppression level.

Images