Thermal runaway prevention method and vehicle

Abstract

The application relates to the technical field of vehicle safety protection, and provides a thermal runaway prevention method and a vehicle. The thermal runaway prevention method comprises the following steps: obtaining running information of a battery pack and environmental information of a location where the vehicle is located. According to the running information, a preset evaluation model is used to determine a current thermal runaway risk index of the battery pack, and a preset prediction model is used to predict a thermal runaway probability of the battery pack in a preset future time period. According to the environmental information, a current prevention demand of the vehicle is determined, and based on the thermal runaway probability, the thermal runaway risk index and the prevention demand, a thermal runaway prevention measure is determined, and the vehicle is controlled to execute the thermal runaway prevention measure. The thermal runaway prevention method can not only identify the possibility of thermal runaway in advance by calculating the thermal runaway risk index and the thermal runaway probability, but also comprehensively determine the thermal runaway prevention measure in combination with the current and future situations, so that the reliability of the battery pack thermal runaway prevention is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle safety protection technology, specifically to a thermal runaway prevention method and a vehicle. Background Technology

[0002] As new energy vehicles equipped with battery packs become more and more common, the battery packs in these vehicles are prone to thermal runaway due to internal short circuits, lithium plating, and diaphragm damage during use.

[0003] Thermal runaway refers to a chain reaction process in which the internal reaction of a battery pack intensifies rapidly, the temperature rises rapidly, a large amount of gas is produced, and the internal pressure increases sharply. Once thermal runaway occurs, if corresponding measures are not taken in time, it can easily lead to serious safety accidents such as fire and explosion. Therefore, thermal runaway monitoring and protection for battery packs are necessary.

[0004] In related technologies, thermal runaway monitoring and protection often involves monitoring the voltage, temperature, and other state parameters of the battery pack. When these parameters exceed corresponding thresholds, thermal runaway intervention is initiated. However, this method can only detect anomalies after thermal runaway has already occurred, and can only passively respond to thermal runaway that has already taken place, resulting in low reliability of thermal runaway protection. Summary of the Invention

[0005] In view of this, this application aims to propose a thermal runaway prevention method to improve the reliability of thermal runaway protection for vehicle battery packs.

[0006] To achieve the above objectives, the technical solution of this application is implemented as follows:

[0007] A thermal runaway prevention method is applied to a vehicle with a battery pack. The method includes: acquiring operational information of the battery pack and acquiring environmental information of the vehicle's location; determining the current thermal runaway risk index of the battery pack using a preset evaluation model based on the operational information, and predicting the probability of thermal runaway of the battery pack within a preset future time period using a preset prediction model; determining the current prevention requirements of the vehicle based on the environmental information, and determining thermal runaway prevention measures based on the thermal runaway probability, the thermal runaway risk index, and the prevention requirements, and controlling the vehicle to execute the thermal runaway prevention measures.

[0008] Furthermore, determining thermal runaway prevention measures based on the thermal runaway probability, the thermal runaway risk index, and the prevention requirements includes: determining the thermal runaway prevention measures according to the thermal runaway risk index when the prevention requirements are preset rapid prevention requirements; and determining the thermal runaway prevention measures according to the thermal runaway probability when the prevention requirements are preset conventional prevention requirements.

[0009] Furthermore, determining the thermal runaway prevention measures based on the thermal runaway risk index includes: determining the current temperature change rate of the battery pack based on the operating information; determining the thermal runaway prevention measures as preset high-level prevention measures if the temperature change rate is not lower than a preset temperature change threshold; and determining the thermal runaway prevention measures based on the prevention level range of the thermal runaway risk index if the temperature change rate is lower than the preset temperature change threshold.

[0010] Furthermore, determining the current thermal runaway risk index of the battery pack based on the operational information and using a preset evaluation model includes: determining the current risk scoring characteristic parameters of the battery pack based on the operational information, wherein the risk scoring characteristic parameters include at least two of the following: temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate; determining the risk impact weight of each of the risk scoring characteristic parameters; and determining the current thermal runaway risk index of the battery pack based on the risk impact weight and the risk scoring characteristic parameters.

[0011] Furthermore, determining the risk impact weight of each of the risk scoring feature parameters includes: determining the baseline impact weight corresponding to each of the risk scoring feature parameters; obtaining the current health status parameters of the battery pack and the current operating condition parameters of the vehicle; and adjusting the baseline impact weight corresponding to each of the risk scoring feature parameters according to the health status parameters and the operating condition parameters to obtain the risk impact weight of each of the risk scoring feature parameters.

[0012] Furthermore, adjusting the baseline influence weights corresponding to each of the risk score feature parameters based on the health status parameters and the operating condition parameters includes: making a preliminary adjustment to the baseline influence weights corresponding to each of the risk score feature parameters based on the health status parameters and the operating condition parameters; obtaining thermal runaway influence parameters through the preset prediction model; determining, based on the thermal runaway influence parameters, whether a secondary adjustment of the initially adjusted baseline influence weights is needed; and, if a secondary adjustment is needed, making a secondary adjustment of the initially adjusted baseline influence weights according to preset weight adjustment rules.

[0013] Furthermore, the preset prediction model adopts the digital twin model of the battery pack.

[0014] Furthermore, the digital twin model includes a thermal-electric coupling physical model and a data prediction model; wherein, the step of using the preset prediction model to predict the probability of thermal runaway of the battery pack within a preset future time period includes: using the thermal-electric coupling physical model to calculate the predicted state data of the battery pack within the preset future time period; inputting the operating information and the predicted state data into the data prediction model, and using the data prediction model to predict the probability of thermal runaway of the battery pack within the preset future time period.

[0015] Furthermore, predicting the probability of thermal runaway of the battery pack within the preset future time period using the data prediction model includes: estimating the offset between the predicted state data obtained from the thermo-electric coupling physical model and the actual state data based on the operating information and the predicted state data; correcting the predicted state data according to the offset; and predicting the probability of thermal runaway of the battery pack within the preset future time period based on the corrected predicted state data and the operating information using the data prediction model.

[0016] Compared with related technologies, this application has at least the following advantages.

[0017] (1) The thermal runaway prevention method described in this application uses a preset evaluation model to evaluate the current thermal runaway risk index of the battery pack. On the other hand, it uses a preset prediction model to predict the probability of thermal runaway of the battery pack in a preset future time period (i.e., thermal runaway probability). Based on the current thermal runaway risk index and the future thermal runaway probability, combined with the prevention requirements, the thermal runaway prevention measures that need to be taken are determined.

[0018] By using the current thermal runaway risk index, we can respond and handle the situation promptly when thermal runaway risks occur, and by using the future thermal runaway probability, we can respond and handle the situation in advance when there is a greater possibility of thermal runaway in the future.

[0019] Therefore, this method not only enables early monitoring and prevention of battery pack thermal runaway, but also avoids the problem of difficulty in taking targeted measures in advance when determining preventive measures solely based on the current thermal runaway risk index, and also avoids the problem of difficulty in responding in a timely manner when thermal runaway problems exist when determining preventive measures solely based on the future thermal runaway probability. In other words, the thermal runaway prevention method of this application comprehensively determines thermal runaway prevention measures by combining the current thermal runaway risk index and the future thermal runaway probability, thereby improving the reliability of thermal runaway protection.

[0020] (2) In the thermal runaway prevention method of this application, since the thermal runaway risk index is calculated directly using the current operating information of the battery pack, while the thermal runaway probability needs to be predicted based on the current operating information of the battery pack to predict the future state before the future thermal runaway probability can be predicted, the calculation speed of the thermal runaway risk index is significantly faster than that of the thermal runaway probability. That is, when the two are calculated in parallel, the thermal runaway risk index will be calculated before the thermal runaway probability. In this application, in scenarios requiring rapid prevention (when the prevention requirement is a preset rapid prevention requirement), the thermal runaway risk index calculated in advance is used to determine the thermal runaway prevention measures, without having to wait for the calculation result of the thermal runaway probability at every moment before determining them. In this way, if thermal runaway occurs at the current moment, since the prevention measures are determined using the pre-calculated thermal runaway risk index, a rapid response can be achieved when thermal runaway anomalies occur. In scenarios where rapid prevention is not required (when the prevention requirement is a preset conventional prevention requirement), the thermal runaway prevention measures are determined using the thermal runaway probability of a preset future time period, which can prevent thermal runaway of the battery pack in advance when it may occur in the future.

[0021] (3) The thermal runaway prevention method of this application also compares the temperature change rate with the preset temperature change threshold. When the temperature change rate is not lower than the preset temperature change threshold (indicating that the current thermal runaway situation is severe), the preset high-level prevention measures are directly triggered for protection. This can minimize the speed of thermal runaway situation expansion and thus improve the reliability of thermal runaway prevention.

[0022] (4) The thermal runaway prevention method of this application also uses at least two of the following parameters as risk scoring characteristic parameters: temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate, to calculate and determine the thermal runaway risk index. This is equivalent to a comprehensive analysis and identification of the thermal runaway risk index from multiple perspectives such as heat, electricity, structure, gas, and pressure. This avoids the limitations of identifying the thermal runaway risk index with a single parameter, thereby improving the accuracy of determining the thermal runaway risk index.

[0023] (5) The thermal runaway prevention method of this application further adjusts the baseline influence weights based on health status parameters and operating condition parameters to obtain the risk influence weights corresponding to each risk score characteristic parameter. This can improve the fit between the risk influence weights and the actual health status and actual operating conditions of the battery, thereby further improving the accuracy of the determination of the thermal runaway risk index.

[0024] (6) The thermal runaway prevention method of this application, when determining the risk impact weight, not only uses the health status and operating conditions of the battery pack, but also uses a preset prediction model to obtain thermal runaway impact parameters, and makes a secondary adjustment to the benchmark impact weight based on the thermal runaway impact parameters. In this way, the determined risk impact weight not only matches the current health status and operating conditions of the battery pack, but also matches the future thermal runaway trend. This advance adjustment of the risk impact weight of the corresponding risk scoring characteristic parameters and advance adaptation to possible future thermal runaway abnormal trends is conducive to improving the calculation accuracy of the thermal runaway risk index.

[0025] (7) The thermal runaway prevention method of this application also utilizes a digital twin model as a preset prediction model. Since the digital twin model can construct a dynamic mapping model that corresponds one-to-one with the physical battery pack in virtual space, and can synchronize all the operating information of the battery pack in real time, and replicate the internal reaction process of the battery pack based on electrochemical and thermodynamic mechanisms, it can deduce the state changes of the battery pack in a preset future time period and predict the evolution trend of thermal runaway in advance. Therefore, the probability of thermal runaway of the output battery pack in a preset future time period can be predicted through this digital twin model, thereby realizing the early prevention of thermal runaway.

[0026] (8) The thermal runaway prevention method of this application also utilizes a thermoelectric coupling physical model to calculate the predicted state data of the battery pack within a preset future time period; inputs the operating information and predicted state data into a data prediction model; and uses the data prediction model to predict the probability of thermal runaway. Since the thermoelectric coupling physical model is built based on electrochemical and thermodynamic theories, it can deduce the future state of the battery from a mechanistic perspective, ensuring that the predicted state data does not deviate from the actual situation. Thus, the thermal runaway probability predicted by the data prediction model based on this predicted state data is more in line with the actual situation, thereby improving the accuracy of the thermal runaway probability prediction.

[0027] (9) The thermal runaway prevention method of this application further corrects the predicted state data before determining the thermal runaway probability when the data prediction model predicts the thermal runaway probability based on the predicted state data. This can reduce the deviation between the predicted state data output by the thermoelectric coupling physical model constructed based on ideal assumptions and the actual state data, thereby reducing the prediction error of the thermal runaway probability and improving the accuracy of the thermal runaway probability prediction.

[0028] Another object of this application is to provide a vehicle equipped with a battery pack, and the vehicle includes a processor and a memory; the memory is used to store a computer program, and the processor executes the computer program to implement the above-described thermal runaway prevention method.

[0029] The vehicle described in this application not only enables early monitoring and prevention of battery pack thermal runaway, but also avoids the problem of difficulty in taking targeted measures in advance when preventive measures are determined solely by the current thermal runaway risk index, and also avoids the problem of difficulty in responding in a timely manner when thermal runaway problems exist when preventive measures are determined solely by the future thermal runaway probability. In other words, the thermal runaway prevention method of this application determines thermal runaway prevention measures by jointly determining the current thermal runaway risk index and the future thermal runaway probability, thereby improving the reliability of thermal runaway protection. Attached Figure Description

[0030] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0031] Figure 1 This is a schematic flowchart of the thermal runaway prevention method described in the embodiments of this application.

[0032] Figure 2 This is a schematic diagram of the process for determining the thermal runaway risk index in the thermal runaway prevention method described in the embodiments of this application.

[0033] Figure 3 This is a schematic diagram of the process for determining the risk impact weight in the thermal runaway prevention method described in the embodiments of this application.

[0034] Figure 4 This is a schematic diagram of the overall process of predicting the probability of thermal runaway in the thermal runaway prevention method described in the embodiments of this application.

[0035] Figure 5 This is a schematic diagram illustrating the specific process of predicting the probability of thermal runaway in the thermal runaway prevention method described in the embodiments of this application. Detailed Implementation

[0036] To make the technical solution and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0037] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0038] Furthermore, it should be noted in the description of this application that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0039] In this application, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0040] The present application will now be described in detail through exemplary embodiments. However, it should be understood that, without further description, elements, structures, and features in one embodiment may be advantageously incorporated into other embodiments.

[0041] An embodiment of the first aspect of this application provides a thermal runaway prevention method applicable to vehicles with battery packs, such as pure electric vehicles, hybrid vehicles, range-extended electric vehicles, etc.

[0042] In vehicles equipped with battery packs, the battery pack typically serves as the vehicle's power supply unit, providing electrical energy to the vehicle's drive system, control system, and onboard electrical equipment. The operational safety and stability of the battery pack directly affect the vehicle's driving reliability and the personal and property safety of the occupants.

[0043] Battery packs may experience thermal runaway during use. Battery thermal runaway refers to a chain reaction process caused by factors such as internal short circuits, lithium plating, separator damage, electrolyte decomposition, overcharging, and over-discharging, leading to a rapid increase in internal reaction intensity, temperature, gas production, and internal pressure. Furthermore, once a battery pack experiences thermal runaway, it may also cause serious safety accidents such as fire, explosion, and jetting.

[0044] For vehicles equipped with battery packs, the internal battery packs often face complex operating conditions during actual operation, such as fast charging, high load, high and low temperature environments, aging and degradation, and mechanical vibration. These conditions make the battery packs more susceptible to thermal runaway. Therefore, thermal runaway monitoring, early warning, and protection are necessary for the battery packs, i.e., thermal runaway prevention. Most related technologies monitor the battery pack's voltage, temperature, and other state parameters to determine if the voltage is below a preset voltage limit and the temperature exceeds a preset temperature limit. This helps determine if thermal runaway has already occurred, and only when thermal runaway is detected are appropriate post-runaway measures taken.

[0045] However, this method can only monitor whether thermal runaway has occurred in the current state. It means that the abnormality can only be detected after the battery pack has already experienced thermal runaway, making it difficult to achieve early warning and protection against thermal runaway.

[0046] In view of this, in order to overcome the shortcomings of related technologies, the thermal runaway prevention method in this embodiment combines... Figure 1 In terms of overall design, it includes the following steps S110-S130.

[0047] Step S110: Obtain the operating information of the battery pack and the environmental information of the vehicle's location.

[0048] The operational information includes operating parameters generated by the battery pack during real-time operation that directly characterize its electrical performance, thermal performance, and internal physical state. For example, operational information includes: battery pack voltage parameters (e.g., the voltage of each individual battery cell), battery pack temperature parameters (e.g., the temperature distribution of the battery pack), battery pack gas parameters (e.g., gas concentration), battery pack cell pressure and deformation parameters (e.g., the internal gas pressure on the cell surface and the deformation of the battery pack), and the battery pack's current SOC (State of Charge) and SOH (State of Health).

[0049] The temperature parameters of the battery pack can be obtained by temperature monitoring components configured inside the battery pack. Specifically, the temperature monitoring components include contact temperature sensors (such as NTC / PT100) attached to the surface of the battery cells or between modules of the battery pack, infrared non-contact temperature sensors for monitoring the temperature distribution of the battery pack casing, and distributed fiber optic temperature sensors embedded inside the battery cells or between modules, such as DTS (Distributed Temperature Sensing).

[0050] The voltage parameters, current state of charge, and current health status of the battery pack can be obtained from the battery management system, which can measure the real-time operating parameters of each individual battery cell within the battery pack, such as the voltage of each individual battery cell.

[0051] The gas parameters of the battery pack can be obtained by a gas monitoring component configured within the battery pack. Specifically, this gas monitoring component may include a miniature gas sensor array arranged in the exhaust channel at the top of the battery module, near the battery pack's pressure relief valve, to monitor the concentration of combustible gases (such as hydrogen, carbon monoxide, and volatile organic compounds, VOCs) released by the battery pack. The gas monitoring component may also include a gas pressure sensor to measure the gas pressure within the battery pack.

[0052] The cell pressure and deformation parameters of the battery pack can be obtained by a pressure deformation monitoring component, which may include, for example, strain gauges attached to the side of the cell (detecting the deformation of the side of the cell and outputting the deformation), and pressure sensors for monitoring the pressure exerted on the casing by the gas inside the cell.

[0053] In addition, this environmental information refers to a data set consisting of external natural environmental parameters and vehicle operating condition parameters. For example, this environmental information includes ambient temperature, ambient humidity, current vehicle speed, charging rate, vehicle altitude, and vehicle load power. Specifically, this environmental information can be obtained by: collecting ambient temperature and humidity data through environmental temperature and humidity sensors installed in the vehicle; obtaining vehicle speed and vehicle load power through the vehicle controller; obtaining the charging rate through the battery management system; and obtaining the vehicle altitude through the vehicle positioning module combined with cloud map data, etc., which will not be elaborated further here.

[0054] Step S120: Based on the operating information, use a preset evaluation model to determine the current thermal runaway risk index of the battery pack, and use a preset prediction model to predict the probability of thermal runaway of the battery pack in a preset future time period.

[0055] The preset evaluation model is a risk assessment calculation model set up to calculate the current thermal runaway risk index of the battery pack.

[0056] The risk assessment calculation model can specifically be: using multiple operating parameters in the operating information, calculating the risk score corresponding to each operating parameter, and then calculating the thermal runaway risk index by weighted averaging of the risk scores of each operating parameter.

[0057] The thermal runaway risk index is a quantitative score calculated using a pre-defined evaluation model based on real-time operating information of the battery pack. It represents the degree of potential thermal runaway risk faced by the battery pack in the current state. For example, the score ranges from 0 to 1, with a value closer to 1 indicating a greater risk of thermal runaway in the battery pack.

[0058] The preset future time period is a pre-defined short-term prediction window used to predict the short-term thermal runaway evolution trend of the battery pack. Its value is determined based on the thermal runaway evolution rate of the battery pack, the vehicle's operating condition response sensitivity, and the sensor sampling frequency. For example, the shortest time from the precursor (such as abnormal temperature rise rate) to the occurrence of thermal runaway is 5-10 seconds, and the longest time does not exceed 60 seconds. Therefore, the duration of the preset future time period can be 5-60 seconds, for example, 10 seconds. This preset future time period represents the period from the current moment to the next 10 seconds.

[0059] The thermal runaway probability is a statistical probability value output by a preset prediction model based on the current operating information of the battery pack, predicting the future thermal runaway development trend of the battery pack. It is used to represent the likelihood of the battery pack experiencing thermal runaway within a preset future time period. Its value range can be, for example, 0%-100%. The higher the value, the greater the probability of thermal runaway within the preset future time period, that is, the greater the possibility of thermal runaway occurring within the preset future time period.

[0060] That is, in step S120, the current thermal runaway risk index of the battery pack is calculated (determined by directly using the current operating information) and the future thermal runaway probability of the battery pack is predicted, so as to realize the current status monitoring and future trend prediction of the thermal runaway risk index of the battery pack, and then determine the corresponding thermal runaway prevention measures in step S130 based on the thermal runaway risk index and thermal runaway probability.

[0061] Step S130: Based on the environmental information, determine the current prevention needs of the vehicle, and based on the thermal runaway probability, thermal runaway risk index, and prevention needs, determine thermal runaway prevention measures, and control the vehicle to implement thermal runaway prevention measures.

[0062] Specifically, in step S130, the current prevention needs of the vehicle are first determined based on environmental information.

[0063] This prevention requirement represents the required thermal runaway prevention response speed for the current vehicle scenario, and is used to characterize the urgency of the thermal runaway warning response currently required by the vehicle. This prevention requirement can be divided into preset rapid prevention requirements and preset routine prevention requirements.

[0064] The pre-defined rapid prevention requirements correspond to scenarios where vehicles are in high-temperature environments, fast charging, high loads, or parked in congested traffic, requiring a rapid response. In these scenarios, once thermal runaway occurs, the situation will deteriorate rapidly within a short period of time, thus necessitating rapid prevention, such as implementing corresponding thermal runaway prevention measures within tens of milliseconds.

[0065] The preset routine prevention requirements correspond to scenarios where the vehicle is at normal temperature, charging slowly, and driving steadily, where the thermal runaway risk index changes steadily and evolves slowly. In these scenarios, the thermal runaway situation usually develops relatively gradually, and once thermal runaway occurs, the degree of thermal runaway will not deteriorate rapidly in a short period of time. Therefore, thermal runaway prevention can be carried out according to the preset routine prevention requirements (such as completing the corresponding thermal runaway prevention measures within a few seconds).

[0066] More specifically, in step S130, prevention requirements can be determined based on parameters such as ambient temperature, charging rate, and load power in the environmental information.

[0067] Specifically, when the ambient temperature exceeds a preset temperature threshold, the charging rate exceeds a preset rate threshold, or the vehicle is under high load, a preset rapid prevention requirement is identified.

[0068] When the ambient temperature is less than or equal to a preset temperature threshold, the charging rate is less than or equal to a preset rate threshold, and the vehicle is under low load or stationary, it is determined as a preset routine prevention requirement.

[0069] Thus, after determining the prevention requirements suitable for the current scenario in step S130, thermal runaway prevention measures can be determined and implemented.

[0070] Thermal runaway prevention measures refer to the early warning, control and protection actions taken to suppress or avoid thermal runaway. Specifically, they can include multi-level prevention measures. The higher the level of the prevention measures, the greater the intervention intensity on the thermal runaway risk index and the better the suppression effect on the thermal runaway risk index.

[0071] For example, this thermal runaway prevention measure includes four levels of prevention measures. The first level of prevention measures refers to the measures taken when the battery is in a normal state and no intervention is required, and the corresponding intervention intensity is the lowest. A first level of prevention measure is, for example, simply maintaining routine monitoring without performing other control intervention actions.

[0072] Level 2 preventative measures refer to measures taken when the battery pack has a low probability of thermal runaway, and the intensity of these interventions is higher than that of Level 1 preventative measures. These Level 2 preventative measures may include enhanced monitoring (e.g., shortening the monitoring interval from the usual 1 second / time to 0.5 seconds / time, or increasing the monitoring frequency of key parameters (e.g., voltage and temperature)), without implementing other control interventions, and without affecting the normal charging and discharging of the battery pack or the normal operation of the vehicle.

[0073] Level 3 preventative measures refer to actions taken when the battery pack exhibits obvious abnormal conditions and has a moderate likelihood of thermal runaway. These measures involve a higher level of intervention than Level 2 preventative measures. For example, Level 3 preventative measures include activating the battery pack's cooling system, initiating cooling circulation, preparing to disconnect the battery pack's high-voltage circuit, and simultaneously sending a warning to the occupants.

[0074] The cooling system may include: a liquid cooling device (for direct or indirect cooling), a spray cooling device (with a liquid storage tank and an evaporation core), and a phase change material (PCM) auxiliary heat absorption device.

[0075] Level 4 preventative measures refer to those taken when the battery pack is in a severe abnormal state and the possibility of thermal runaway is extremely high. These measures represent the highest level of intervention and are designed to minimize the hazards of thermal runaway. For example, Level 4 preventative measures include activating the battery pack cooling system and maximizing the coolant flow rate (specifically, the flow rate of the liquid cooling device and the spray cooling device) to rapidly reduce the battery pack temperature; immediately executing fault isolation actions to prevent the fault from spreading; activating the battery pack's built-in fire extinguishing device; turning on the vehicle's ventilation and smoke extraction system; triggering the vehicle's audible and visual alarms; sending rescue information to the cloud platform; and alerting the occupants to evacuate.

[0076] Specifically, the fault cut-off action can be to cut off the high-voltage circuit of the vehicle battery pack.

[0077] Furthermore, when determining the thermal runaway prevention measures that need to be taken, the early warning measures that need to be implemented can be determined from the multi-level prevention measures based on the thermal runaway risk index, thermal runaway probability, and prevention needs.

[0078] Specifically, the higher the thermal runaway risk index of a battery pack, the higher the probability of short-term thermal runaway, the higher the degree of danger, the higher the intensity of intervention required for thermal runaway prevention measures, and the more necessary it is to select high-level prevention measures.

[0079] In addition, the higher the predicted probability of thermal runaway of the battery pack, the greater the possibility of thermal runaway in the future. Therefore, the intervention intensity of thermal runaway prevention measures is also higher, and higher-level prevention measures need to be selected.

[0080] Furthermore, prevention requirements characterize the abnormal response speed needed for the current vehicle scenario: for preset rapid prevention requirements, a faster response strategy is needed; for preset routine prevention requirements, a stable response strategy can be adopted. Therefore, when determining thermal runaway prevention measures, it is also necessary to consider the abnormal response speed required for the current scenario.

[0081] Therefore, through steps S110-S130, on the one hand, the current thermal runaway risk index level of the battery pack is evaluated using a preset evaluation model, and on the other hand, the probability of thermal runaway of the battery pack occurring within a preset future time period (i.e., thermal runaway probability) is predicted using a preset prediction model. Based on the current thermal runaway risk index level and the future thermal runaway probability, combined with prevention requirements, the thermal runaway prevention measures that need to be taken are jointly determined.

[0082] By using the current thermal runaway risk index level, we can respond and handle the situation promptly when thermal runaway risk occurs, and by using the future thermal runaway probability, we can respond and handle the situation in advance when there is a greater possibility of thermal runaway in the future.

[0083] Therefore, this method not only enables early monitoring and prevention of battery pack thermal runaway, but also avoids the problem of difficulty in taking targeted measures in advance when determining preventive measures solely based on the current thermal runaway risk index, and also avoids the problem of difficulty in responding in a timely manner when thermal runaway problems exist when determining preventive measures solely based on the future thermal runaway probability. In other words, the thermal runaway prevention method of this application determines thermal runaway prevention measures by jointly determining the current thermal runaway risk index and the future thermal runaway probability, which can improve the reliability of thermal runaway protection.

[0084] Continue to refer to Figure 1 and combined Figure 2As shown, in some exemplary embodiments, in step S120 above, the current thermal runaway risk index of the current battery pack is determined based on the operating information and using a preset evaluation model, which may specifically include the following steps S121-S123.

[0085] Step S121: Determine the current risk score characteristic parameters of the battery pack based on the operating information.

[0086] Among them, the risk scoring characteristic parameters refer to parameters that can reflect whether the battery pack has thermal runaway, and can specifically include at least two of the following: temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate.

[0087] Specifically, the temperature rise rate refers to the change in the average surface temperature of the battery pack per unit time. The formula is: Temperature rise rate = (Current temperature - Temperature at the previous sampling time) / Sampling time interval. A higher temperature rise rate indicates more intense heat generation in the battery and a higher risk of thermal runaway.

[0088] The voltage descent rate refers to the rate of voltage drop within a unit of time. It is calculated as: Voltage Descent Rate = (Lowest Single Cell Voltage at the Previous Sampling Time - Current Lowest Single Cell Voltage) / Sampling Time Interval. When a battery pack experiences thermal runaway, the voltage drops rapidly due to internal abnormalities. Therefore, a higher voltage descent rate indicates a more severe internal abnormality and a higher risk of thermal runaway.

[0089] Local temperature difference refers to the temperature difference between different cells or regions within a battery pack (it can be the difference between the maximum and minimum temperatures within the battery pack). When thermal runaway occurs in a battery pack, some cells or regions will initially show abnormal temperature increases, while other normal cells or regions will experience gradual temperature changes. This leads to a larger temperature difference between different cells or regions. Therefore, a larger local temperature difference indicates poorer cell consistency, more pronounced local anomalies, and a higher risk of thermal runaway.

[0090] The rate of change of combustible gas concentration refers to the increase in the concentration of combustible gas inside the battery pack per unit time. The formula is: Rate of change of combustible gas concentration = (Current combustible gas concentration - Combustible gas concentration at the previous sampling time) / Sampling time interval. Combustible gases refer to flammable gases such as hydrogen, methane, ethylene, and carbon monoxide generated inside the battery pack when thermal runaway occurs.

[0091] When thermal runaway occurs inside a battery, flammable gases may be released inside the battery, and the separator of the battery pack may be damaged. Due to the damage to the separator and the gas release inside the battery, the concentration of flammable gases inside the battery pack will rise rapidly. Therefore, the rate of change of flammable gas concentration will be greater, and the greater the rate of change of flammable gas concentration, the more serious the abnormality inside the battery and the greater the risk index of thermal runaway of the battery pack.

[0092] The pressure change rate refers to the rate of increase in internal pressure of a battery pack per unit time. It is calculated as: Pressure Change Rate = (Current Internal Pressure - Internal Pressure at the Previous Sampling Time) / Sampling Time Interval. When a battery pack experiences thermal runaway, a large amount of flammable gas is generated internally, causing a rapid increase in internal pressure. Therefore, a higher pressure change rate indicates more severe gas production within the battery pack, and thus a higher risk of thermal runaway.

[0093] Specifically, in step S121, at least two parameters can be selected from the rate of temperature rise, rate of voltage drop, local temperature difference, rate of change of combustible gas concentration, and rate of change of pressure as risk scoring characteristic parameters. For example, at least two parameters can be arbitrarily selected, such as the rate of temperature rise and the rate of voltage drop. Alternatively, a combination of temperature-related, voltage-related, and gas-related parameters can be considered, selecting one parameter of each type, for a total of three parameters as risk scoring characteristic parameters.

[0094] In addition, in some embodiments, all five parameters mentioned above (temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate) can be selected as risk scoring characteristic parameters. For ease of explanation, the following explanation will use temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate as risk scoring characteristic parameters.

[0095] Step S122: Determine the risk impact weights of each risk scoring characteristic parameter.

[0096] Specifically, the sum of the weights of each risk impact is 1, and the higher the risk impact weight of the risk scoring characteristic parameter, the stronger the indicative effect of the risk scoring characteristic parameter on thermal runaway.

[0097] Different risk scoring characteristic parameters exhibit varying degrees of timeliness, accuracy, and intuitiveness in their manifestation of thermal runaway. Some parameters reflect thermal runaway more quickly and accurately, demonstrating a stronger indicative role and thus carrying a higher risk impact weight. Other parameters, however, are less timely and accurate in indicating thermal runaway (e.g., their manifestation is similar to that of other anomalies). These parameters have a weaker indicative role and a lower risk impact weight.

[0098] Therefore, in step S122, it is necessary to determine the risk impact weight of each risk scoring parameter so that the current thermal runaway risk index of the battery pack can be calculated in the subsequent step S123.

[0099] In some embodiments, refer to Figure 3 In step S122, the risk impact weight of each risk scoring characteristic parameter is determined, which may specifically include the following steps S1221-S1223.

[0100] Step S1221: Determine the benchmark influence weights corresponding to each risk score characteristic parameter.

[0101] Among them, the baseline influence weight refers to the pre-set initial weight. Specifically, the baseline influence weight can be pre-set by staff based on experiments and experience. The specific rules for setting the baseline influence weight include: the more accurate the evaluation of thermal runaway state, the stronger the characterization ability, and the higher the early warning reliability (equivalent to the stronger the indication effect of thermal runaway), the greater the corresponding baseline influence weight. The sum of the baseline influence weights corresponding to all risk scoring characteristic parameters is 1.

[0102] More specifically, for the rate of temperature rise, rate of voltage drop, local temperature difference, rate of change of combustible gas concentration, and rate of change of pressure, the rate of temperature rise is the earliest and most direct manifestation of thermal runaway. From the precursors of thermal runaway (minor internal short circuits, minor side reactions) to the occurrence of thermal runaway, the rate of temperature rise can always accurately capture abnormal changes inside the battery, unaffected by other external factors (such as battery pack sealing, environmental interference). Therefore, the rate of temperature rise can most accurately evaluate the real-time state of thermal runaway, has the highest early warning reliability, and the strongest indication of thermal runaway, thus having the largest benchmark influence weight.

[0103] Regarding the voltage descent rate, it is the core manifestation of abnormal battery pack performance during the development of thermal runaway. However, the voltage descent occurs later than the temperature rise rate (usually, a significant voltage descent will only occur after the temperature rise rate becomes abnormal), and it is easily affected by external factors such as fluctuations in charging and discharging current. Therefore, its timeliness and accuracy in reflecting the current manifestation of thermal runaway are weaker than those of the temperature rise rate. Consequently, its baseline influence weight is less than that of the temperature rise rate.

[0104] As for local temperature difference, it is a manifestation of uneven temperature distribution during the development of thermal runaway. When some cells experience significant temperature rise (the rate of temperature rise is abnormal) and form a temperature difference with other cells, obvious abnormalities will appear. Its timeliness and intuitiveness in reflecting the current manifestation of thermal runaway are weaker than the voltage drop rate, and its accuracy in evaluating the thermal runaway state is weaker than the voltage drop rate. Therefore, its benchmark influence weight ranks third.

[0105] Regarding the rate of change of combustible gas concentration, the increase in combustible gas concentration is a manifestation of chemical anomalies during the development of thermal runaway. However, its evaluation accuracy is weaker than that of local temperature difference due to the influence of battery pack sealing performance. Therefore, its benchmark influence weight ranks fourth.

[0106] As for the rate of pressure change, the pressure rise is a clear manifestation of the middle and late stages of thermal runaway, and an indirect manifestation of the large-scale production of combustible gas. It appears the latest and is easily affected by external factors such as battery pack sealing performance and exhaust system status. Therefore, it has the weakest accuracy in evaluating the thermal runaway state and thus has the smallest baseline influence weight.

[0107] For example, the baseline influence weight of the temperature rise rate can be 0.4, the baseline influence weight of the voltage drop rate can be 0.25, the baseline influence weight of the local temperature difference can be 0.15, the baseline influence weight of the combustible gas concentration change rate can be 0.12, and the baseline influence weight of the pressure change rate can be 0.08. The sum of the above baseline influence weights is 1.

[0108] Step S1222: Obtain the current health status parameters of the battery pack and the current operating condition parameters of the vehicle.

[0109] Among them, the health status parameter refers to the core parameters that can reflect the degree of aging and performance degradation of the battery pack. Specifically, it can include: SOH (State of Health, which is the ratio of the battery's current actual capacity to its rated capacity).

[0110] Specifically, battery packs with a high degree of aging tend to have a higher thermal runaway risk index for the same temperature fluctuation compared to battery packs with a low degree of aging (in other words, the same temperature fluctuation can have different levels of danger). Therefore, when calculating the thermal runaway risk index, the greater the degree of aging of the battery pack, the greater the baseline influence weight of its temperature rise rate should be.

[0111] Operating condition parameters refer to core parameters that reflect the current workload of the vehicle and the load status of the battery pack. These can include: charge / discharge rate, ambient temperature, vehicle speed, etc.

[0112] Regarding operating parameters, the risk of thermal runaway differs for the same voltage jump in high-temperature, fast-charging scenarios compared to normal-temperature, slow-charging scenarios. The thermal runaway risk index is higher in high-temperature fast-charging scenarios. Therefore, in high-temperature fast-charging scenarios, it is necessary to increase the risk impact weight of the risk scoring characteristic parameters reflecting voltage jumps, that is, to increase the weight of the voltage descent rate.

[0113] Therefore, in step S1222, the health status parameters and operating condition parameters of the battery pack are first obtained. Then, in step S1223, the risk impact weights of each risk scoring characteristic parameter are determined based on the health status parameters, operating condition parameters, and the influence weights of each benchmark.

[0114] Step S1223: Based on the health status parameters and operating condition parameters, adjust the baseline influence weights corresponding to each risk score characteristic parameter to obtain the risk influence weights of each risk score characteristic parameter.

[0115] Specifically, in step S1223, the adjustment rules for adjusting the baseline influence weights based on health status parameters and operating condition parameters include:

[0116] First, for the health status parameter, when SOH < preset health status value (e.g., 80%), a first preset weight (e.g., 0.1) can be added to the baseline influence weight corresponding to the temperature rise rate, while the risk influence weights corresponding to other risk scoring characteristic parameters remain unchanged.

[0117] Secondly, for the operating condition parameters, first determine the current scenario of the vehicle, and then adjust the weight of each benchmark influence according to the weight adjustment rules adapted to the current scenario of the vehicle.

[0118] Specifically, the vehicle's charging scenarios can be categorized into four types: high-temperature fast charging, normal-temperature fast charging, high-temperature slow charging, and normal-temperature slow charging. Each scenario corresponds to a weighting adjustment rule.

[0119] Specifically, when the ambient temperature exceeds a preset temperature threshold (e.g., 45℃) and the charging rate exceeds a preset rate threshold (the preset rate threshold is a criterion for distinguishing between fast and slow charging conditions, used to help determine the vehicle's current preventative needs; its value is determined based on the battery pack's charge / discharge rate tolerance, cell material characteristics, and vehicle charging system design parameters; for example, the upper limit of fast charging rate tolerance for mainstream vehicle power batteries is 2C, so the preset rate threshold can be set to 2 coulombs), the vehicle is identified as operating in a high-temperature fast charging scenario. In a high-temperature fast charging scenario, battery heat generation is intense, and the risks of temperature and voltage anomalies are significantly increased. Therefore, the weighting rule for this high-temperature fast charging scenario can be: increasing the baseline influence weight of the temperature rise rate by a second preset weight (e.g., 0.12), increasing the baseline influence weight of the voltage drop rate by a third preset weight (e.g., 0.1), while keeping the risk influence weights for other risk scoring characteristic parameters unchanged.

[0120] When the ambient temperature is less than or equal to a preset temperature threshold and the charging rate exceeds a preset rate threshold, the vehicle is identified as operating in a normal temperature fast charging scenario. In this scenario, battery voltage fluctuation risks are significant. Therefore, the weighting rule for this high-temperature fast charging scenario can be: increase the baseline impact weight of the voltage descent rate by a fourth preset weight (less than the third preset weight, for example, 0.05), while keeping the risk impact weights for other risk scoring feature parameters unchanged.

[0121] When the ambient temperature exceeds a preset temperature threshold and the charging rate is less than or equal to a preset rate threshold, the vehicle is identified as operating in a high-temperature slow charging scenario. In this scenario, battery heat generation is intense, significantly increasing the risk of temperature anomalies. Therefore, the weighting rule for this high-temperature slow charging scenario can be: increase the baseline influence weight of the temperature rise rate by a fifth preset weight (less than the second preset weight, for example, 0.1), while keeping the risk influence weights of other risk scoring feature parameters unchanged.

[0122] When the ambient temperature is less than or equal to a preset temperature threshold and the charging rate is less than or equal to a preset rate threshold, the vehicle is identified as operating in a normal temperature slow charging scenario. In a normal temperature slow charging scenario, the operating conditions are stable, and the corresponding operating condition weighting rule can be: the risk impact weights corresponding to each risk scoring feature parameter remain unchanged.

[0123] It is worth noting that if the weight of a certain risk scoring feature parameter becomes negative after adjustment, the weight will be reset to the preset minimum weight threshold (e.g., 0.01) to avoid calculation errors caused by negative weights.

[0124] Furthermore, in step S1223, after adjusting the weights of each benchmark influence according to the adjustment rules to obtain the adjusted benchmark influence weights, in order to ensure that the sum of the weights is 1, it is also necessary to normalize the adjusted benchmark influence weights (divide all the adjusted benchmark influence weights by the sum of the benchmark influence weights). The normalized benchmark influence weights can then be used as the risk influence weights corresponding to the risk scoring feature parameters.

[0125] Thus, through steps S1221-S1223, based on the baseline influence weights, the influence weights of each baseline are adjusted according to the health status parameters and operating condition parameters to obtain the risk influence weights corresponding to each risk scoring characteristic parameter. This can improve the fit between the risk influence weights and the actual health status and actual operating conditions of the battery, thereby further improving the accuracy of determining the thermal runaway risk index.

[0126] Furthermore, when adjusting the benchmark impact weights corresponding to each risk score characteristic parameter, one can also consider adjusting the current benchmark impact weights based on the anomalies that may occur in the future time period.

[0127] That is, in some embodiments, step S1223 above adjusts the baseline influence weights corresponding to each risk score characteristic parameter based on health status parameters and operating condition parameters, specifically including:

[0128] First, based on the health status parameters and operating condition parameters, the baseline influence weights corresponding to each risk score characteristic parameter are initially adjusted.

[0129] The adjustment method for making preliminary adjustments to the influence weights of each benchmark can refer to the adjustment rules in the embodiment corresponding to step S1223 above, and will not be repeated here.

[0130] Next, thermal runaway impact parameters are obtained through a preset prediction model. Based on these parameters, it is determined whether a secondary adjustment is needed to the initially adjusted baseline impact weights. If a secondary adjustment is required, the baseline impact weights are adjusted according to preset weight adjustment rules.

[0131] Among them, thermal runaway impact parameters refer to the core parameters that can characterize the development trend of thermal runaway within a preset future time period. Specifically, they can include the peak temperature rise rate and the voltage anomaly probability within the preset future time period.

[0132] Among them, the peak temperature rise rate refers to the maximum temperature rise rate of the battery pack within a preset future time period (e.g., 30s), and the unit is ℃ / s. The voltage anomaly probability refers to the percentage of times within a preset future time period when the voltage drop rate of a single cell is greater than or equal to a preset anomaly threshold (e.g., 0.1V / s), and the value ranges from 0% to 100%.

[0133] Specifically, when the peak temperature rise rate exceeds the preset temperature threshold, it indicates that the temperature rise is predicted to be faster and the temperature rise anomaly is more obvious in the preset future time period. Therefore, when the peak temperature rise rate exceeds the preset temperature threshold, it indicates that the baseline influence weight after the initial adjustment needs to be adjusted again.

[0134] In addition, when the voltage anomaly probability exceeds the preset anomaly probability threshold, it indicates that the probability of voltage anomaly is high in the preset future time period, and the voltage fluctuation is obvious. The accuracy of determining the thermal runaway risk index by using the voltage descent rate is higher. Therefore, when the voltage anomaly probability exceeds the preset anomaly probability threshold, it indicates that the baseline influence weight after the initial adjustment needs to be adjusted again.

[0135] Furthermore, the preset weight adjustment rules during the secondary adjustment include:

[0136] When the peak temperature rise rate exceeds the preset temperature threshold, a sixth preset weight (with a value of 0.05-0.1, for example, 0.8) is added to the initial adjusted baseline influence weight corresponding to the temperature rise rate to obtain the secondary adjusted baseline influence weight corresponding to that temperature rise rate. The baseline influence weights of other risk scoring characteristic parameters remain unchanged.

[0137] When the voltage anomaly probability exceeds the preset anomaly probability threshold, a seventh preset weight (the value can be determined by the staff based on experience, for example, the value range can be 0.05-0.1, for example, 0.8) is added to the initial adjusted baseline influence weight corresponding to the voltage descent rate, to obtain the secondary adjusted baseline influence weight corresponding to the voltage descent rate. The baseline influence weights of other risk scoring characteristic parameters remain unchanged.

[0138] It is worth noting that when the peak temperature rise rate exceeds the preset temperature threshold and the voltage anomaly probability exceeds the preset anomaly probability threshold, a sixth preset weight is added to the initial adjusted baseline influence weight corresponding to the temperature rise rate, and a seventh preset weight is added to the initial adjusted baseline influence weight corresponding to the voltage drop rate. The baseline influence weights of other risk scoring characteristic parameters remain unchanged.

[0139] Furthermore, the peak temperature rise rate and the voltage anomaly probability within the preset future time period can be predicted using the aforementioned preset prediction model. Specifically, in predicting the probability of thermal runaway, the preset prediction model first predicts the state data within the preset future time period based on battery operating data. It then obtains the corresponding temperature and voltage parameters from this predicted state data. By processing these temperature and voltage parameters accordingly, the peak temperature rise rate and the voltage anomaly probability can be obtained, thus yielding the parameters influencing thermal runaway.

[0140] Furthermore, after adjusting the basic influence weights of each risk scoring characteristic parameter a second time, it is necessary to normalize the adjusted basic influence weights to obtain the risk influence weights corresponding to each risk scoring characteristic parameter.

[0141] Therefore, when determining the risk impact weights, not only are they determined based on the battery pack's health status and operating conditions, but also by using a pre-set prediction model to obtain thermal runaway impact parameters. The baseline impact weights are then adjusted based on these parameters. This ensures that the determined risk impact weights not only match the current battery pack's health status and operating conditions but also anticipate future thermal runaway trends. This advance adjustment of the risk impact weights for the corresponding risk scoring characteristic parameters, adapting to potential future thermal runaway anomalies, helps improve the accuracy of the thermal runaway risk index calculation.

[0142] In addition, after determining the various risk scoring characteristic parameters and their corresponding risk impact weights in steps S121 and S122, step S123 can be executed to determine the current thermal runaway risk index of the battery pack.

[0143] Step S123: Determine the current thermal runaway risk index of the battery pack based on the risk impact weight and risk scoring characteristic parameters.

[0144] Specifically, when determining the current thermal runaway risk index, the original detection values ​​of each risk scoring feature parameter are first processed according to a preset standardization conversion rule, converting them into standardized values ​​between 0 and 1. Then, the standardized values ​​of each risk scoring feature parameter are multiplied by their corresponding risk impact weights. Finally, the products of all risk scoring feature parameters are summed to obtain the thermal runaway risk index score (i.e., the current thermal runaway risk index).

[0145] Specifically, the pre-defined standardized conversion rule may include, for example, the standardized value = (original detection value - minimum parameter value) / (maximum parameter value - minimum parameter value). Here, the minimum and maximum parameter values ​​are the critical range values ​​of each risk score characteristic parameter obtained in advance by staff through numerous thermal runaway experiments, at the time of thermal runaway.

[0146] Thus, by using at least two of the following parameters as risk scoring characteristic parameters through steps S121-S123 above—temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate—to calculate and determine the thermal runaway risk index, it is equivalent to comprehensively analyzing and identifying the thermal runaway risk index from multiple perspectives such as heat, electricity, structure, gas, and pressure. This avoids the limitations of identifying the thermal runaway risk index with a single parameter, thereby improving the accuracy of determining the thermal runaway risk index.

[0147] Continuing from the above Figures 1-3 As shown, in some of the exemplary embodiments, the aforementioned preset prediction model employs a digital twin model of the battery pack.

[0148] A digital twin model is a virtual mapping model built in virtual space based on the physical and thermodynamic characteristics of a battery pack. It corresponds one-to-one with the physical battery pack and can map the current operating state of the physical battery pack in real time and predict the future operating state of the battery pack.

[0149] In this embodiment, a digital twin model is used as the preset prediction model. Because the digital twin model can construct a dynamic mapping model in virtual space that corresponds one-to-one with the physical battery pack, and can synchronize all operational information of the battery pack in real time, and based on electrochemical and thermodynamic mechanisms, replicate the internal reaction process of the battery pack, and deduce the state changes of the battery pack within a preset future time period, it can predict the evolution trend of thermal runaway in advance. Therefore, this digital twin model can predict the probability of thermal runaway of the battery pack within a preset future time period, thereby achieving early prevention of thermal runaway.

[0150] Continuing from the above Figures 1 to 3and combined Figure 4 As shown, in some exemplary embodiments, the digital twin model is used to predict the probability of thermal runaway within a preset future time period, which may specifically include a thermal-electric coupling physical model and a data prediction model. Specifically, in step S120, based on operational information, the preset prediction model is used to predict the probability of thermal runaway of the battery pack within a preset future time period, which may include steps S410-S420 below.

[0151] Step S410: Calculate the predicted state data of the battery pack in a preset future time period using a thermo-electric coupling physical model.

[0152] Among them, the predicted state data refers to the operating information of the battery pack at each moment within a preset future time period, which includes the operating parameters of the battery pack at each moment, such as the voltage parameters of the battery pack (e.g., the voltage of each battery cell), the temperature parameters of the battery pack (e.g., the temperature distribution of the battery pack), the gas parameters of the battery pack (e.g., gas concentration), the cell pressure and deformation parameters of the battery pack (e.g., the internal gas pressure on the cell surface and the deformation of the battery pack), and the current SOC, SOH and other operating parameters of the battery pack.

[0153] Specifically, the thermo-electric coupling physical model is a battery pack physical model built based on electrochemical and thermodynamic theories. It is used to calculate the heat generation and electrochemical state in a preset future time period based on the current battery operation data and using electrochemical and thermal equations, thereby obtaining the operation information at each moment in the preset future time period, that is, the predicted state data.

[0154] Specifically, this thermo-electric coupling physical model essentially simulates the interaction process of "electrochemical reaction generating heat and heat accumulation affecting electrochemical reaction" during battery operation. The core is to calculate the real-time electrical performance parameters (such as electric potential and current density) and thermal performance parameters (such as temperature and heat generation rate) of the battery by conforming to actual physical laws, and predict their future change trends to obtain a model of predicted state data.

[0155] The thermo-electric coupling physical model can specifically be a simplified P2D model (Pseudo Two-Dimensional Model).

[0156] Specifically, the thermo-electric coupling physical model includes a set of mathematical equations describing the electrochemical reactions, mass transfer, charge conservation, and energy conservation within the battery. These equations may include: the electrochemical equation of state, the charge conservation equation, the energy conservation equation, and the heat generation equation.

[0157] The electrochemical equation of state characterizes the diffusion of lithium ions, the rate of electrochemical reactions, and the changes in solid / liquid potential within the battery, simulating the movement of lithium ions in the electrodes and electrolyte and their reactions to generate current. This electrochemical equation of state can specifically be the Butler-Volmer equation (also known as the charge transfer kinetics equation), which will not be elaborated upon here.

[0158] The charge conservation equation is used to characterize the relationship between the internal current distribution and charge balance of a battery.

[0159] The heat generation equation is used to characterize the heat generated during battery operation, that is, to simulate the total heat generated by irreversible reactions (such as resistance heating) and reversible reactions (such as enthalpy change) during electrochemical reactions. For example: Qtotal = Qohm + Qpol + Qrxn. Where Qohm is the ohmic heat, Qohm = I 2 R0 and Qpol are the heat of polarization, which is the energy dissipation caused by electrochemical polarization (activation overpotential η), and Qpol = Iη; Qrxn is the heat of reaction, which is the entropy change heat of the electrochemical reaction itself and the heat of side reactions.

[0160] The energy conservation equation describes the transfer and accumulation of heat inside a battery, simulating the process by which generated heat is conducted inside the battery, exchanged with the outside environment, and ultimately leads to a change in battery temperature. Specifically, the energy conservation equation states that within a unit volume, heat accumulation = heat conduction + internal heat generation. This includes: .

[0161] Where T is temperature (K or °C); ρ is the density of the battery material (kg / m³); Cp is the specific heat capacity (J / (kg·K)); and k is the thermal conductivity (W / (m·K)). Qtotal is the heat generation rate (W / m³) at the corresponding time, which is equivalent to the heat generation at the current time, and its formula is the heat generation rate equation below.

[0162] In step S410, the real-time collected current operating information of the battery pack is input into the thermo-electric coupling physical model. The thermo-electric coupling physical model combines the built-in electrochemical equation, heat conduction equation and heat generation rate equation to calculate the current heat generation rate, temperature field distribution and electrochemical state of the battery pack.

[0163] Then, the thermo-electric coupling physical model iteratively extrapolates along the time dimension based on the current state. Under the assumption that the future charging current remains unchanged and the ambient temperature remains stable, it extrapolates and calculates the battery temperature, SOC, gas production and electrochemical evolution trend within a preset future time period, and outputs predicted state data.

[0164] Step S420: Input the operating information and predicted status data into the data prediction model, and predict the probability of thermal runaway of the battery pack in a preset future time period through the data prediction model.

[0165] Specifically, the data prediction model can adopt a lightweight temporal neural network model, such as an LSTM+Attention model (long short-term memory network + attention mechanism model). The operating information and predicted state data are input into the data prediction model, and the data prediction model can output the probability of thermal runaway of the battery within a preset future time period.

[0166] Specifically, refer to Figure 5 In some exemplary embodiments, step S420, which involves predicting the probability of thermal runaway of the battery pack within a preset future time period using a data prediction model, may specifically include steps S421-S423 below.

[0167] Specifically, the data prediction model may include a first task processing submodule, an overlay module, and a second task processing submodule.

[0168] The first task processing submodule employs a lightweight temporal neural network model to estimate the offset of the predicted state data in step S421 below. The overlay module is used in step S422 below to correct the predicted state data based on the offset, obtaining the corrected predicted state data. The second task processing submodule also employs a lightweight temporal neural network model to predict the probability of thermal runaway of the battery pack within a preset future time period in step S423 below.

[0169] Step S421: Based on the operating information and predicted state data, use the data prediction model to estimate the offset between the predicted state data obtained from the thermal-electric coupling physical model and the actual state data.

[0170] Specifically, the thermo-electric coupling physical model is based on theoretically calculated predicted state data, and its construction relies on idealized electrochemical and thermodynamic assumptions. However, the actual operation of a battery is affected by a variety of complex factors, such as battery aging, ambient temperature fluctuations, unstable charging and discharging currents, sensor measurement errors, and differences in battery pack sealing performance. Therefore, the predicted state data obtained using the thermo-electric coupling physical model will have a significant deviation from the actual future state data.

[0171] Therefore, in step S421, the offset between the predicted state data and the actual state data obtained from the thermo-electric coupling physical model is first estimated, and then in step S422, the predicted state data is corrected based on the offset.

[0172] When the predicted state data includes multiple operating parameters, the offset includes multiple parameter offsets, with one parameter offset corresponding to each operating parameter. For example, if the predicted state data includes three parameters: cell temperature, single cell voltage, and pack pressure, then the offsets will include temperature offset, voltage offset, and pressure offset.

[0173] More specifically, in step S421, the offset can be predicted by the first task processing submodule mentioned above. The first task processing submodule can be a lightweight temporal neural network model (with an internal core architecture of LSTM+Attention) and is pre-trained with a large amount of sample data, including multiple historical sample data.

[0174] Each historical sample data set includes: a) a sequence of historical operational information, i.e., the historical operational information of the battery pack at various points in time within a specific historical period; b) a sequence of physical model predictions, i.e., the predicted state data for a predetermined future time period obtained by the thermo-electric coupling physical model at each point in time; c) the offset of the battery pack at each point in time within that historical period (manually labeled). Additionally, each historical sample data set may also include: d) a sequence of ambient temperature changes, i.e., the ambient temperature within that historical period. It is worth noting that if the influence of ambient temperature on the offset is ignored, the ambient temperature change sequence may not be included in the historical sample data sets.

[0175] In this way, the first task processing submodule can learn the dynamic change trajectory of the running information in various historical sample data. This dynamic change trajectory represents the comprehensive state of the battery pack, including its aging state, health state, and changes in sealing performance. Furthermore, through a large amount of historical sample data, the first task processing submodule can learn the evolution of the correspondence between the predicted state data and the offset of the thermo-electric coupling physical model under each comprehensive state of the battery pack and under different ambient temperature changes.

[0176] After training is completed, the predicted state data and operation information are input into the first task processing submodule. The environmental temperature change records over a recent period can also be input into the first task processing submodule. The first task processing submodule can then analyze the dynamic change trajectory of the battery pack's operation information based on the battery pack's operation information over a historical period, and analyze the current environmental temperature change based on the environmental temperature change records.

[0177] Then, the first task processing submodule determines the offset corresponding to the current battery pack based on the dynamic change trajectory learned from historical sample data, the evolution law of the correspondence between the predicted state data and the offset under the corresponding combination of environmental temperature changes, and the current predicted state data.

[0178] Step S422: Correct the predicted state data according to the offset.

[0179] Specifically, in step S422, after the first task processing submodule outputs the offset, the superposition module corrects the predicted state data output by the thermo-electric coupling physical model based on the offset, obtaining the corrected predicted state data. Specifically, after the first task processing submodule outputs the offset to the superposition module, the superposition module superimposes the offset (each running parameter is superimposed with its corresponding parameter offset) onto the predicted state data to obtain the corrected predicted state data.

[0180] After correcting the predicted state data to obtain the corrected predicted state data, the following step S423 is performed to predict the probability of thermal runaway based on the corrected predicted state data.

[0181] Step S423: Based on the corrected predicted state data and operating information, use the data prediction model to predict the probability of thermal runaway of the battery pack within a preset future time period.

[0182] Specifically, in step S423, the second task processing submodule in the data prediction model is used to predict the probability of thermal runaway and output the probability of thermal runaway of the battery in a preset future time period.

[0183] The second task processing submodule also uses the LSTM+Attention model, which is a lightweight temporal neural network model. In other embodiments, classification or regression models can also be used, such as multilayer fully connected networks (MLP).

[0184] In this embodiment, the second task processing submodule performs feature fusion between the corrected predicted state data and the current operating information, and outputs a probability value of 0-100%. The higher the probability value, the greater the possibility of thermal runaway in the future.

[0185] Specifically, the second task processing submodule is pre-trained based on a large amount of sample data. The training process is as follows: First, sufficient sample data is collected, including scenarios such as normal charging and discharging, accelerated aging, and thermal runaway triggering in bench tests, as well as historical thermal runaway probability sample data under actual vehicle driving and charging scenarios. The process of obtaining this thermal runaway probability prediction sample data includes: staff statistically analyzing the historical thermal runaway probability (the percentage of times thermal runaway occurred in the corresponding combination of operational information and future state data) for each combined scenario. This historical thermal runaway probability is then annotated in the operational information and future state data to obtain the thermal runaway probability prediction sample data for the second processing sub-model to learn from.

[0186] Secondly, the historical thermal runaway probability sample data is divided into a training set (80%), a validation set (10%), and a test set (10%). The training set data is input into the initial LSTM+Attention model, and the error between the predicted thermal runaway probability and the actual labeled result is used as the loss function. The model parameters (such as the number of hidden layer neurons in the LSTM and the weight coefficients of the Attention mechanism) are iteratively adjusted through the backpropagation algorithm. Then, the model performance is verified using the validation set data. Finally, the final performance of the model is tested using the test set data to ensure that the model prediction error is controlled within a preset range (such as error ≤ 5%), thus completing the training.

[0187] The corrected predicted state data and current operating information are input into the second task processing submodule of the data prediction model, which can then calculate the probability of thermal runaway.

[0188] Therefore, through steps S421-S423, when the data prediction model predicts the probability of thermal runaway based on the predicted state data, the predicted state data is first corrected before the thermal runaway probability is determined. This can reduce the deviation between the predicted state data output by the thermoelectric coupling physical model constructed based on ideal assumptions and the actual state data, thereby reducing the prediction error of the thermal runaway probability and improving the accuracy of the thermal runaway probability prediction.

[0189] Furthermore, through steps S410-S420, the predicted state data of the battery pack within a preset future time period is calculated using a thermoelectric coupling physical model; the operating information and predicted state data are input into the data prediction model; and the probability of thermal runaway is predicted through the data prediction model. Since the thermoelectric coupling physical model is built upon electrochemical and thermodynamic theories, it can deduce the future state of the battery from a mechanistic perspective, ensuring that the predicted state data does not deviate from the actual situation. Therefore, the thermal runaway probability predicted by the data prediction model based on this predicted state data is more closely aligned with the actual situation, thereby improving the accuracy of the thermal runaway probability prediction.

[0190] In addition, in some embodiments, the initial parameters of the digital twin model (such as the thermal conductivity and heat generation rate coefficient of the thermo-electric coupling physical model, and the weight coefficient of the data prediction model) are set based on idealized experimental conditions. As the battery ages and environmental conditions change during use, the initial parameters will gradually become disconnected from the actual operating state of the battery, resulting in a decrease in the accuracy of model prediction.

[0191] Therefore, in this embodiment, the current thermal runaway situation can also be determined based on the currently calculated thermal runaway risk index, so as to adjust the relevant parameters of the digital twin model and thereby improve the prediction accuracy of the digital twin model.

[0192] Specifically, the adjustment methods for adjusting the parameters of the digital twin model based on this thermal runaway risk index include:

[0193] First, the thermal conductivity k in the thermo-electric coupling physical model can be adjusted based on the thermal runaway risk index. Thermal conductivity k is a core parameter describing the rate of heat transfer within the battery, and its value directly affects the calculation results of the heat conduction equation, thus impacting the accuracy of temperature-related parameters in the predicted state data. As the battery ages, the internal material structure of the cell changes, and its thermal conductivity gradually decreases. Simultaneously, when the thermal runaway risk index increases, the rate of heat generation inside the battery accelerates, and the heat transfer pattern differs from that under normal conditions. If the thermal conductivity k remains fixed, it will lead to increased temperature prediction deviation, thereby affecting the accuracy of thermal runaway probability prediction.

[0194] Therefore, the thermal conductivity K can be adjusted according to the thermal runaway risk index. For example, the adjustment method could be as follows: When the thermal runaway risk index score is ≥0.6 (indicating medium or higher risk), it indicates a high probability of thermal runaway occurring inside the battery, and the heat transfer rate is lower than normal. In this case, the current value of the thermal conductivity k should be reduced by 5%-10% (the specific reduction ratio is dynamically adjusted based on the risk score; the higher the risk score, the larger the reduction ratio, such as 10% reduction when the risk score is ≥0.8, and 5% reduction when 0.6≤risk score<0.8); when the thermal runaway risk index score is <0.3 (extremely low risk), it indicates that the battery is operating normally and its thermal conductivity is stable, and the thermal conductivity k should be restored to its initial baseline value; when 0.3≤risk score<0.6 (lower risk), the current value of the thermal conductivity k should remain unchanged. This makes the heat transfer calculation of the thermo-electric coupling physical model more consistent with the actual situation of the battery, reduces the deviation of temperature-related predicted state data, provides more accurate temperature data support for subsequent thermal runaway probability prediction, and improves the model's ability to capture the precursors of thermal runaway.

[0195] Secondly, the Attention weight coefficients in the data prediction model can be adjusted. These coefficients are used to assign attention weights to each input parameter, and their core function is to make the model focus on parameters that have a significant impact on thermal runaway (such as the rate of temperature rise and the rate of voltage drop). Because the level of the thermal runaway risk index directly reflects the degree of influence of different parameters on thermal runaway, when the thermal runaway risk index increases, the influence of parameters such as the rate of temperature rise and the rate of voltage drop will be significantly enhanced. If the Attention weight coefficients remain fixed, it will be impossible to dynamically focus on these key parameters, which will lead to a decrease in the model's sensitivity to identifying thermal runaway anomalies.

[0196] Therefore, the Attention weight coefficients can be adjusted based on the thermal runaway risk index. For example, the adjustment method could be: extract the two risk score feature parameters with the highest weight in the thermal runaway risk index calculation (such as the rate of temperature rise and the rate of voltage drop); when the thermal runaway risk index score is ≥0.6, increase the Attention weight coefficients corresponding to these two parameters by 8%-12%, while decreasing the Attention weight coefficients of other parameters (such as the rate of pressure change) by 3%-5%, ensuring the model focuses on high-impact parameters; when the thermal runaway risk index score is <0.3, restore the Attention weight coefficients of all parameters to their initial training values; when 0.3 ≤ risk score <0.6, only increase the Attention weight coefficients of the core parameters by 3%-5%. This can improve the model's attention to key precursor parameters of thermal runaway, enhance the model's sensitivity to identifying thermal runaway anomalies, and reduce the prediction bias of thermal runaway probability. Especially when the thermal runaway risk index increases, it can more quickly and accurately capture the thermal runaway evolution trend.

[0197] The thermal runaway probability and thermal runaway risk index can be determined through the above embodiments. After determining the thermal runaway probability and thermal runaway risk index, the above step S130 is executed to determine thermal runaway prevention measures based on the thermal runaway probability, thermal runaway risk index, and prevention requirements.

[0198] Continue to refer to Figure 1 In some exemplary implementations, step S130 above, based on the probability of thermal runaway, the thermal runaway risk index, and prevention requirements, determines thermal runaway prevention measures, which may specifically include:

[0199] When the prevention requirement is a preset rapid prevention requirement, thermal runaway prevention measures are determined based on the thermal runaway risk index. When the prevention requirement is a preset conventional prevention requirement, thermal runaway prevention measures are determined based on the thermal runaway probability.

[0200] Specifically, under the premise of a pre-defined rapid prevention requirement, this means that once thermal runaway occurs in the current scenario, the situation will rapidly deteriorate within a short period of time. For example, in high-temperature fast charging scenarios, the battery generates heat rapidly, and thermal runaway can occur from warning signs to the actual event in just a few seconds to tens of seconds. If a rapid response is not taken, it can easily lead to serious safety accidents such as battery fires and explosions. Therefore, it is necessary to quickly identify abnormal thermal runaway situations and implement relevant thermal runaway prevention measures in a timely manner upon detection to achieve a rapid response.

[0201] The calculation of the thermal runaway risk index is simpler, consumes less computing power, and is faster than the prediction of the thermal runaway probability. Specifically, the thermal runaway risk index can be quickly calculated based on current operating information through parameter extraction, standardization, and weighted summation, with a computation time of less than 10ms. In contrast, the thermal runaway probability requires deduction through a thermo-electric coupling physical model and correction through a data prediction model, resulting in a relatively longer computation time (approximately 50-100ms).

[0202] Therefore, in this embodiment, the steps of calculating the thermal runaway risk index and predicting the thermal runaway probability are executed in parallel. Furthermore, when the prevention requirement is a preset rapid prevention requirement, the thermal runaway risk index can be calculated first after it is completed. Since the thermal runaway risk index is calculated quickly, thermal runaway prevention measures can be determined first based on the thermal runaway risk index without waiting for the predicted thermal runaway probability. This allows for a rapid response and avoids protection delays caused by waiting for the predicted thermal runaway probability.

[0203] When the prevention requirements are the preset routine prevention requirements, the response speed requirement for thermal runaway prevention is relatively low. For example, in scenarios such as slow charging at room temperature and stable driving, the development of thermal runaway is gradual, and it usually takes several minutes from the warning signs to the occurrence. Therefore, we can wait for the prediction result of the thermal runaway probability (i.e., the predicted thermal runaway probability) and determine the thermal runaway prevention measures based on the thermal runaway probability in the preset future time period.

[0204] Specifically, when determining thermal runaway prevention measures based on the probability of thermal runaway, it may include: pre-setting four levels of thermal runaway probability, namely, extremely low probability range (e.g., thermal runaway probability < 30%), low probability range (e.g., 30% ≤ thermal runaway probability < 60%), high probability range (e.g., 60% ≤ thermal runaway probability < 80%), and extremely high probability range (e.g., thermal runaway probability ≥ 80%).

[0205] Each probability interval corresponds to a preset thermal runaway prevention measure. After obtaining the thermal runaway probability, the level interval of that probability is first determined, and then the thermal runaway prevention measure to be executed is determined according to the correspondence between the probability interval and the thermal runaway prevention measure. For example, when the thermal runaway probability is in the extremely low probability interval, the first-level prevention measure (routine monitoring only) is executed; when it is in the low probability interval, the second-level prevention measure (enhanced monitoring) is executed; when it is in the high probability interval, the third-level prevention measure (cooling intervention + early warning) is executed; and when it is in the extremely high probability interval, the fourth-level prevention measure (emergency shutdown + fire fighting and rescue) is executed.

[0206] Thus, in scenarios requiring rapid prevention (where the prevention requirement is a preset rapid prevention requirement), thermal runaway prevention measures can be determined by calculating a faster thermal runaway risk index, enabling rapid response in the event of thermal runaway anomalies. In scenarios where rapid prevention is not required (where the prevention requirement is a preset conventional prevention requirement), thermal runaway prevention measures can be determined using the thermal runaway probability over a preset future time period, allowing for proactive prevention when the battery pack may experience thermal runaway in the future.

[0207] It is worth noting that, when the prevention requirement is the preset conventional prevention requirement, this embodiment can directly determine the thermal runaway prevention measures using only the thermal runaway probability. If the current thermal runaway risk index is low (corresponding to a low level of prevention measures) and the thermal runaway probability is high (corresponding to a high level of prevention measures), it means that although the possibility of thermal runaway occurring at present is low, it is highly likely to occur in the future. Therefore, determining the thermal runaway prevention measures using the thermal runaway probability can achieve early prevention.

[0208] If the current thermal runaway risk index is high and the thermal runaway probability is low, it means that although the possibility of thermal runaway is high at present, the probability of thermal runaway in the future is low. This indicates that there may be transient disturbances at present, but there is no great possibility of thermal runaway in the future. Therefore, thermal runaway prevention measures can be determined by using the thermal runaway probability.

[0209] Similarly, when the level of preventive measures corresponding to the thermal runaway risk index and the thermal runaway probability is the same, the thermal runaway probability can be used to determine the thermal runaway preventive measures.

[0210] In some other embodiments, when the prevention requirement is a preset conventional prevention requirement, the thermal runaway probability can be used as the primary factor, and the thermal runaway risk index as an auxiliary factor, with both used together to determine the thermal runaway prevention measures. In this case, the step of determining whether the level of the prevention measure corresponding to the thermal runaway risk index is less than or equal to the level of the prevention measure corresponding to the thermal runaway probability can be performed first. If it is less than or equal to, the thermal runaway prevention measures can be directly determined using the probability range to which the thermal runaway probability belongs. If the level of the prevention measure corresponding to the thermal runaway risk index is greater than the level of the prevention measure corresponding to the thermal runaway probability, the prevention measure corresponding to the higher-level thermal runaway risk index can be used.

[0211] In some embodiments, when the prevention requirement is a preset rapid prevention requirement, determining the thermal runaway prevention measures based on the thermal runaway risk index may specifically include: determining the current temperature change rate of the battery pack based on operational information; if the temperature change rate is not lower than a preset temperature change threshold, determining the thermal runaway prevention measures as preset high-level prevention measures; if the temperature change rate is lower than the preset temperature change threshold, determining the thermal runaway prevention measures based on the prevention level range of the thermal runaway risk index.

[0212] Specifically, the rate of temperature change is equivalent to the rate of temperature rise in the above embodiments.

[0213] The preset high-level preventive measures can be either the third-level preventive measures in the above embodiments or the fourth-level preventive measures. There is no limitation here. The specific measures can be set according to the actual needs of the scenario. For example, in the high-temperature fast charging scenario, the preset high-level preventive measures are the fourth-level preventive measures, and in the high-load driving scenario, the preset high-level preventive measures are the third-level preventive measures.

[0214] More specifically, if the rate of temperature change is not lower than a preset temperature change threshold, it indicates that the battery pack is highly likely to experience thermal runaway. At this point, the battery is generating intense heat internally, and the temperature is rising rapidly, requiring immediate and intensive intervention. Therefore, preset high-level preventative measures can be implemented to intervene directly, preventing the thermal runaway from worsening due to delayed intervention, thereby minimizing the risk of safety accidents caused by thermal runaway and protecting the personal and property safety of passengers.

[0215] In addition, if the temperature change rate is lower than the preset temperature change threshold, it means that the battery pack is less likely to experience thermal runaway and the current thermal runaway situation is developing slowly. Therefore, in this case, it is not necessary to directly adopt the preset high-level prevention measures. Instead, wait for the calculation result of the thermal runaway risk index and determine the thermal runaway prevention measures according to the prevention level range to which the thermal runaway risk index belongs.

[0216] Specifically, the prevention level ranges of the thermal runaway risk index are pre-set. Based on a large number of thermal runaway experiments and real vehicle operation data, combined with the quantitative scoring range of the thermal runaway risk index, the staff preset four prevention level ranges: extremely low risk range (score < 0.3), relatively low risk range (0.3 ≤ score < 0.6), medium risk range (0.6 ≤ score < 0.8), and extremely high risk range (score ≥ 0.8). Each prevention level range corresponds to a preset thermal runaway prevention measure. For example, the extremely low risk range corresponds to the first level of prevention measure, the relatively low risk range corresponds to the second level of prevention measure, the medium risk range corresponds to the third level of prevention measure, and the extremely high risk range corresponds to the fourth level of prevention measure. After obtaining the thermal runaway risk index score, the corresponding prevention measure can be directly matched without additional calculation, ensuring a rapid response.

[0217] In addition, the preset temperature change threshold is used to determine whether emergency intervention is needed, and is used to quickly identify whether the battery pack is in a state of impending thermal runaway. The value of the preset temperature change threshold can be preset by the staff based on the type, rated capacity, service life of the battery pack, and actual experimental data. For example, it can be dynamically adjusted according to the current state of health (SOH) parameter of the battery pack. For example, the preset temperature change threshold can usually be 1.5% / s (i.e., the temperature change per second is not less than 1.5% of the current temperature). For battery packs with a high degree of aging (SOH < 80%), the preset temperature change threshold can be lowered to 1.2% / s to improve the sensitivity of emergency state identification; for new batteries (SOH ≥ 95%), the preset temperature change threshold can be raised to 1.8% / s to avoid over-intervention due to misjudgment.

[0218] In this way, by comparing the rate of temperature change with the preset temperature change threshold, if the rate of temperature change is not lower than the preset temperature change threshold, it indicates that the current thermal runaway situation is severe. By directly triggering the preset high-level prevention measures, the speed of thermal runaway expansion can be reduced to the maximum extent, thereby improving the reliability of thermal runaway prevention.

[0219] Furthermore, it's worth noting that when the prevention requirement is a preset rapid prevention requirement, after determining the thermal runaway prevention measures based on the thermal runaway risk index, if the thermal runaway probability is calculated, the currently applied thermal runaway prevention measures can be updated and adjusted based on the thermal runaway probability. For example, after determining the thermal runaway probability, the corresponding prevention measures are determined. Then, from the level of the prevention measures corresponding to that thermal runaway probability and the level of the currently applied prevention measures (i.e., the level of the prevention measures corresponding to the thermal runaway risk index), the higher-level prevention measure is selected, and the currently applied prevention measures are upgraded to the higher-level prevention measures.

[0220] It is worth noting that, regarding the thermal runaway prevention method of this embodiment, based on the above exemplary implementations, in specific implementation, as a preferred embodiment, it is still based on... Figure 1-5 As shown, it may include, for example:

[0221] First, obtain the battery pack's operating information and the environmental information of the vehicle's location.

[0222] Then, based on operational information, the current thermal runaway risk index of the battery pack is calculated and determined. Simultaneously, a digital twin model, incorporating both a thermoelectric coupling physical model and a data prediction model, is used to predict the probability of thermal runaway of the battery pack within a preset future timeframe. Furthermore, current preventative measures for the vehicle are determined based on environmental information.

[0223] Subsequently, when the prevention requirement is a preset rapid prevention requirement, since the thermal runaway risk index is calculated quickly, the corresponding thermal runaway prevention measures are determined and implemented based on the thermal runaway risk index.

[0224] When the prevention requirement is the preset routine prevention requirement, the corresponding thermal runaway prevention measures are determined and implemented based on the probability of thermal runaway.

[0225] When determining thermal runaway prevention measures, multi-level prevention measures are implemented based on the thermal runaway risk index / probability. These measures include activating the cooling system, preparing to perform a fault cutoff action, or increasing the cooling system's flow rate and performing a fault cutoff action to disconnect the high-voltage circuit of the vehicle's battery pack, thereby achieving precise protection against thermal runaway of the battery pack.

[0226] In the preferred embodiment of thermal runaway prevention described above, the specific implementation process of each step can still be found in the descriptions of the above exemplary embodiments, and the beneficial effects of each step in this preferred embodiment can also be found in the descriptions of the above exemplary embodiments.

[0227] The thermal runaway prevention method in this embodiment adopts the above design. On the one hand, it uses a preset evaluation model to evaluate the current thermal runaway risk index of the battery pack. On the other hand, it uses a preset prediction model to predict the probability of thermal runaway of the battery pack in a preset future time period (i.e., thermal runaway probability). Based on the current thermal runaway risk index and the future thermal runaway probability, combined with the prevention requirements, the thermal runaway prevention measures that need to be taken are determined.

[0228] By using the current thermal runaway risk index, we can respond and handle the situation promptly when thermal runaway risks occur, and by using the future thermal runaway probability, we can respond and handle the situation in advance when there is a greater possibility of thermal runaway in the future.

[0229] Therefore, this method not only enables early monitoring and prevention of battery pack thermal runaway, but also avoids the problem of difficulty in taking targeted measures in advance when determining preventive measures solely based on the current thermal runaway risk index, and also avoids the problem of difficulty in responding in a timely manner when thermal runaway problems exist when determining preventive measures solely based on the future thermal runaway probability. In other words, the thermal runaway prevention method of this application comprehensively determines thermal runaway prevention measures by combining the current thermal runaway risk index and the future thermal runaway probability, thereby improving the reliability of thermal runaway protection.

[0230] An embodiment of the second aspect of this application provides a vehicle, more specifically a vehicle equipped with a battery pack. The vehicle specifically includes a processor and a memory.

[0231] The memory stores the application code that executes the solution of this application, and its execution is controlled by the processor. The processor executes the application code stored in the memory to implement the content shown in the foregoing method embodiments.

[0232] The vehicle in this embodiment can not only achieve early monitoring and prevention of battery pack thermal runaway, but also avoid the problem of difficulty in taking targeted measures in advance when determining preventive measures solely based on the current thermal runaway risk index, and the problem of difficulty in responding in a timely manner when thermal runaway problems exist when determining preventive measures solely based on the future thermal runaway probability. In other words, the thermal runaway prevention method of this application determines thermal runaway prevention measures by jointly determining the current thermal runaway risk index and the future thermal runaway probability, thereby improving the reliability of thermal runaway protection.

[0233] The above descriptions are merely some embodiments of this application and are not intended to limit this application. The technical features or structures in the foregoing different embodiments can be arbitrarily combined to form other specific technical solutions as needed. For those skilled in the art, this application can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of the claims of this application.

Claims

1. A method for preventing thermal runaway, applied to vehicles with battery packs, characterized in that, The method includes: Obtain the operating information of the battery pack and the environmental information of the vehicle's location; Based on the operational information, a preset evaluation model is used to determine the current thermal runaway risk index of the battery pack, and a preset prediction model is used to predict the probability of thermal runaway of the battery pack in a preset future time period. Based on the environmental information, the current prevention needs of the vehicle are determined, and based on the thermal runaway probability, the thermal runaway risk index, and the prevention needs, thermal runaway prevention measures are determined, and the vehicle is controlled to execute the thermal runaway prevention measures.

2. The thermal runaway prevention method according to claim 1, characterized in that, The determination of thermal runaway prevention measures based on the thermal runaway probability, the thermal runaway risk index, and the prevention requirements includes: When the prevention requirement is a preset rapid prevention requirement, the thermal runaway prevention measures are determined based on the thermal runaway risk index. When the prevention requirement is a preset conventional prevention requirement, the thermal runaway prevention measures are determined based on the thermal runaway probability.

3. The thermal runaway prevention method according to claim 2, characterized in that, The determination of thermal runaway prevention measures based on the thermal runaway risk index includes: Based on the operational information, determine the current temperature change rate of the battery pack; If the rate of temperature change is not lower than a preset temperature change threshold, the thermal runaway prevention measure is determined to be a preset high-level prevention measure. If the temperature change rate is lower than the preset temperature change threshold, the thermal runaway prevention measures are determined according to the prevention level range of the thermal runaway risk index.

4. The thermal runaway prevention method according to claim 1, characterized in that, The step of determining the current thermal runaway risk index of the battery pack based on the operational information and using a preset evaluation model includes: Based on the operational information, the current risk score characteristic parameters of the battery pack are determined, wherein the risk score characteristic parameters include at least two of the following: temperature rise rate, voltage drop rate, local temperature difference, combustible gas concentration change rate, and pressure change rate. Determine the risk impact weights of each of the aforementioned risk scoring characteristic parameters; The current thermal runaway risk index of the battery pack is determined based on the risk impact weight and the risk scoring characteristic parameters.

5. The thermal runaway prevention method according to claim 4, characterized in that, The determination of the risk impact weights for each of the risk scoring feature parameters includes: Determine the benchmark influence weights corresponding to each of the aforementioned risk scoring feature parameters; Obtain the current health status parameters of the battery pack and the current operating condition parameters of the vehicle; Based on the health status parameters and the operating condition parameters, the baseline influence weights corresponding to each of the risk scoring feature parameters are adjusted to obtain the risk influence weights of each of the risk scoring feature parameters.

6. The thermal runaway prevention method according to claim 5, characterized in that, The step of adjusting the baseline influence weights corresponding to each of the risk score feature parameters based on the health status parameters and the operating condition parameters includes: Based on the health status parameters and the operating condition parameters, the baseline influence weights corresponding to each of the risk score feature parameters are initially adjusted. The thermal runaway impact parameters are obtained through the preset prediction model. Based on the thermal runaway impact parameters, it is determined whether the preliminary adjusted benchmark impact weights need to be adjusted again. If a secondary adjustment is required, the influence weights of each benchmark after the initial adjustment are adjusted according to the preset weight adjustment rules.

7. The method for preventing thermal runaway according to any one of claims 1 to 6, characterized in that: The preset prediction model uses the digital twin model of the battery pack.

8. The thermal runaway prevention method according to claim 7, characterized in that: The digital twin model includes a thermo-electric coupling physical model and a data prediction model; The step of using a preset prediction model to predict the probability of thermal runaway of the battery pack within a preset future time period includes: Using the aforementioned thermo-electric coupling physical model, the predicted state data of the battery pack within the preset future time period is calculated; The operating information and the predicted state data are input into the data prediction model, and the probability of thermal runaway of the battery pack in the preset future time period is predicted by the data prediction model.

9. The thermal runaway prevention method according to claim 8, characterized in that, The step of predicting the probability of thermal runaway of the battery pack within the preset future time period using the data prediction model includes: Based on the operational information and the predicted state data, the offset between the predicted state data obtained from the thermo-electric coupling physical model and the actual state data is estimated using the data prediction model. The predicted state data is corrected based on the offset. Based on the corrected predicted state data and the operational information, the data prediction model is used to predict the probability of thermal runaway of the battery pack within the preset future time period.

10. A vehicle equipped with a battery pack, characterized in that: Including processor and memory; The memory is used to store a computer program, and the processor executes the computer program to implement the thermal runaway prevention method according to any one of claims 1-9.

Images