State switching method, device and system for energy storage equipment in vehicle and vehicle

Abstract

The invention discloses a state switching method, device and system for energy storage equipment in a vehicle and the vehicle. The method comprises the steps that first state information of energy storage equipment in a vehicle and second state information of the environment where the vehicle is located are obtained; based on the first state information and the second state information, state evaluation information of the energy storage equipment at a future moment is predicted, and the future moment is later than the current moment; in response to the state evaluation information that the running state is an abnormal state, the initial control strategy of the vehicle is adjusted through performance adjustment factors of the energy storage equipment, a target control strategy is obtained, and the performance adjustment factors are used for adjusting the running performance of the energy storage equipment; and at the future moment, according to the target control strategy, the vehicle is controlled to switch the running state of the energy storage equipment from the abnormal state to the safe state. The technical problem that the state switching accuracy of the energy storage equipment in the vehicle is low is solved.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and more specifically, to a method, apparatus, system, and vehicle for switching the state of an energy storage device in a vehicle. Background Technology

[0002] Currently, battery thermal management is a key technology for ensuring battery safety and extending battery life.

[0003] In related technologies, battery thermal management systems mostly employ "threshold-triggered" control logic. The basic principle is based on real-time monitoring of battery temperature and voltage; when parameters exceed preset threshold ranges, a cooling or heating mechanism is automatically activated to adjust the battery's operating state. While this control method is simple and direct, its shortcomings become increasingly apparent in complex and rapidly changing driving environments. Therefore, the technical problem of low accuracy in the state switching of energy storage devices in vehicles remains.

[0004] There is currently no effective solution to the aforementioned technical problems. Summary of the Invention

[0005] This invention provides a method, apparatus, system, and vehicle for switching the state of energy storage devices in a vehicle, in order to at least solve the technical problem of low accuracy in switching the state of energy storage devices in a vehicle.

[0006] According to one aspect of the present invention, a method for switching the state of an energy storage device in a vehicle is provided. The method includes: acquiring first state information of the energy storage device in the vehicle and second state information of the environment in which the vehicle is located, wherein the first state information represents the operating state of the energy storage device in the environment at the current moment, and the second state information represents the environmental state affecting the operating state; predicting state assessment information of the energy storage device at a future moment based on the first and second state information, wherein the future moment is later than the current moment; responding to the state assessment information indicating an abnormal operating state, adjusting the initial control strategy of the vehicle using a performance adjustment factor of the energy storage device to obtain a target control strategy, wherein the performance adjustment factor is used to adjust the operating performance of the energy storage device, and the operating performance of the energy storage device controlled according to the target control strategy is greater than the operating performance of the energy storage device controlled according to the initial control strategy, and the target control strategy represents the rule for controlling the vehicle to switch from an abnormal state; and at a future moment, controlling the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state according to the target control strategy.

[0007] Optionally, based on the first state information and the second state information, predicting the state assessment information of the energy storage device at a future time includes: performing noise reduction processing on the first state information and the second state information to obtain the noise-reduced first state information and the noise-reduced second state information, wherein the accuracy of the noise-reduced first state information is greater than the accuracy of the first state information before noise reduction, and the accuracy of the noise-reduced second state information is greater than the accuracy of the second state information before noise reduction; constructing a time-series feature matrix based on the noise-reduced first state information and the noise-reduced second state information according to the time window of the target period; and using the time-series feature matrix to determine the state assessment information.

[0008] Optionally, the state assessment information is determined using a time-series feature matrix, including: inputting the time-series feature matrix into the long short-term memory network model corresponding to the energy storage device, and using the long short-term memory network model to output the state assessment information for a future time after the current time, which includes at least one of the following: distribution information, heating rate, and risk information. The distribution information is used to represent the temperature distribution in the energy storage device, and the risk information is used to represent the risk level of the energy storage device being in an abnormal state.

[0009] Optionally, the performance adjustment factors include at least one of the following: temperature adjustment factors, risk adjustment factors, and energy consumption adjustment factors. The temperature adjustment factor is used to adjust the degree of temperature fluctuation of the energy storage device; the risk adjustment factor is used to adjust the risk level of the energy storage device in an abnormal state; and the energy consumption adjustment information is used to adjust the energy consumption of the energy storage device. In response to the state assessment information indicating that the operating state is abnormal, the initial control strategy of the vehicle is adjusted using the performance adjustment factors of the energy storage device to obtain a target control strategy, including at least one of the following: in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy is adjusted using the deep learning model corresponding to the energy storage device, with the temperature adjustment factor being less than or equal to the temperature fluctuation threshold as the objective, to obtain the target control strategy; in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy is adjusted using the deep learning model, with the risk adjustment factor being less than the risk level threshold as the objective, to obtain the target control strategy; and in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy is adjusted using the deep learning model, with the energy consumption adjustment factor being minimized as the objective, to obtain the target control strategy.

[0010] Optionally, the vehicle includes a cooling system and a heating system. The control strategy includes a cooling strategy and a heating strategy. The cooling strategy represents the rules for performing cooling operations on the energy storage device, and the heating strategy represents the rules for performing heating operations on the energy storage device to adjust for abnormal states. In the future, according to the target control strategy, the vehicle is controlled to switch the operating state of the energy storage device from an abnormal state to a safe state. This includes: in the future, according to the cooling strategy, controlling the cooling system to perform cooling operations on the energy storage device to switch the operating state from an abnormal state to a safe state; and in the future, according to the heating strategy, controlling the heating system to perform heating operations on the energy storage device to switch the operating state from an abnormal state to a safe state.

[0011] Optionally, the vehicle also includes a safety actuator. The control strategy includes a first control strategy and a second control strategy. The first control strategy represents a rule for adjusting the internal pressure of the energy storage device, and the second control strategy represents a rule for disconnecting the power to the energy storage device. In the future, according to the target control strategy, the vehicle is controlled to switch the operating state of the energy storage device from an abnormal state to a safe state. This includes: in the future, according to the first control strategy, controlling the explosion-proof valve of the safety actuator to adjust the internal pressure of the energy storage device to switch the operating state from an abnormal state to a safe state, wherein the adjusted internal pressure of the energy storage device is lower than the internal pressure of the energy storage device before adjustment; and in the future, according to the second control strategy, controlling the relay of the safety actuator to disconnect the circuit between the energy storage device and the electrical system in the vehicle to switch the operating state from an abnormal state to a safe state.

[0012] Optionally, the method further includes: after the energy storage device switches from an abnormal state to a safe state, acquiring the first state information of the energy storage device in the safe state; determining the temperature error of the energy storage device based on the first state information; an adjustment step, in response to the temperature error being greater than an error threshold, determining the future time as the current time, using gradient descent to adjust the model parameters of the deep learning model corresponding to the energy storage device at the current time, and / or the model parameters of the long short-term memory network model of the energy storage device at the current time, to obtain an adjustment result; in response to the adjustment result at the current time being that the temperature error is greater than the error threshold, determining the future time of the target duration after the current time, and returning to execute from the beginning of the adjustment step until the adjustment result is that the temperature error is less than the error threshold.

[0013] According to another aspect of the present invention, a state switching device for an energy storage device in a vehicle is also provided. The device may include: an acquisition unit, configured to acquire first state information of the energy storage device in the vehicle and second state information of the environment in which the vehicle is located, wherein the first state information represents the operating state of the energy storage device in the environment at the current moment, and the second state information represents the environmental state affecting the operating state; a prediction unit, configured to predict state assessment information of the energy storage device at a future moment based on the first and second state information, wherein the future moment is later than the current moment; an adjustment unit, configured to, in response to the state assessment information indicating that the energy storage device is in an abnormal state, adjust the initial control strategy of the vehicle using a performance adjustment factor of the energy storage device to obtain a target control strategy, wherein the performance adjustment factor is used to adjust the operating performance of the energy storage device, and the operating performance of the energy storage device controlled according to the target control strategy is greater than the operating performance of the energy storage device controlled according to the initial control strategy, and the target control strategy represents the rule for controlling the vehicle to switch from an abnormal state; and a control unit, configured to, at a future moment, control the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state according to the target control strategy.

[0014] According to another aspect of the present invention, a state switching system for an energy storage device in a vehicle is also provided. The system may include: a perception layer, configured to acquire first state information of the energy storage device in the vehicle and second state information of the environment in which the vehicle is located, wherein the first state information represents the operating state of the energy storage device in the environment at the current moment, and the second state information represents the environmental state affecting the operating state; a decision layer, configured to predict state assessment information of the energy storage device at a future moment based on the first and second state information, wherein the future moment is later than the current moment; in response to the state assessment information indicating that the energy storage device is in an abnormal state, adjusting the initial control strategy of the vehicle using the performance adjustment factor of the energy storage device to obtain a target control strategy, wherein the performance adjustment factor is used to adjust the operating performance of the energy storage device, and the operating performance of the energy storage device controlled according to the target control strategy is greater than the operating performance of the energy storage device controlled according to the initial control strategy, and the target control strategy represents the rule for controlling the vehicle to switch from an abnormal state; and an execution layer, configured to control the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state at a future moment according to the target control strategy.

[0015] Optionally, the system further includes: a feedback layer, used to acquire first state information of the energy storage device in the safe state after the energy storage device switches from an abnormal state to a safe state; determine the temperature error of the energy storage device based on the first state information; and adjust the model parameters of the deep learning model corresponding to the energy storage device and / or the model parameters of the long short-term memory network model of the energy storage device in response to the temperature error being greater than the error threshold.

[0016] According to another aspect of the present invention, a computer-readable storage medium is also provided. The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the methods described in the embodiments of the present invention.

[0017] According to another aspect of the present invention, a processor is also provided. This processor is used to run a program, wherein the program executes the methods described above in the embodiments of the present invention during runtime.

[0018] According to another aspect of the present invention, an electronic device is also provided. The electronic device includes a memory and a processor, the memory storing a computer program, and the processor being configured to run the computer program to perform the methods described in the embodiments of the present invention.

[0019] According to another aspect of the present invention, a computer program product is also provided. This computer program product includes a computer program that, when executed by a processor, implements the methods described above in the embodiments of the present invention.

[0020] According to another aspect of the present invention, a vehicle is also provided. The vehicle includes a memory and a processor. The memory stores an executable program; the processor runs the program, which, when executed, implements the methods described in the embodiments of the present invention.

[0021] In this embodiment of the invention, by fusing the first state information of the battery with the second state information of the vehicle environment, an advanced predictive algorithm is used to assess the future state of the energy storage device in advance, significantly improving the accuracy of state switching. Specifically, this embodiment introduces real-time monitoring and analysis based on multi-source data, not limited to temperature and voltage, but also covering early thermal runaway indicators such as gas concentration and internal pressure, enabling early warning before the energy storage device enters an abnormal state. By predicting future state assessment information, the performance adjustment factors of the energy storage device, such as dynamically adjusting heating power, cooling efficiency, and safety level, can be proactively utilized before an abnormal state occurs to optimize the initial control strategy and form a target control strategy. The execution of the above-mentioned target control strategy ensures that the energy storage device can be switched from an abnormal state to a safe state in a timely manner in the future, avoiding thermal management failure caused by state judgment delays, thereby effectively solving the problem of low state switching accuracy in related technologies, improving the efficiency of battery thermal management and the overall safety of the vehicle. This solves the technical problem of low state switching accuracy of energy storage devices in vehicles and achieves the technical effect of improving the state switching accuracy of energy storage devices in vehicles. Attached Figure Description

[0022] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0023] Figure 1 This is a flowchart of a state switching method for an energy storage device in a vehicle according to an embodiment of the present invention;

[0024] Figure 2 This is a flowchart of an artificial intelligence-based battery thermal management system and a multimodal main control closed loop according to an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of an artificial intelligence-based battery thermal management system and a multimodal safety control sub-strategy according to an embodiment of the present invention;

[0026] Figure 4 This is a schematic diagram of a thermal insulation control sub-strategy according to an embodiment of the present invention;

[0027] Figure 5(a) is a schematic diagram of a battery system according to an embodiment of the present invention;

[0028] Figure 5(b) is a schematic diagram of another battery system according to an embodiment of the present invention;

[0029] Figure 6(a) is a schematic diagram of the interior of a battery system according to an embodiment of the present invention;

[0030] Figure 6(b) is a schematic diagram of the interior of another battery system according to an embodiment of the present invention;

[0031] Figure 7 This is a schematic diagram of a state switching device for an energy storage device in a vehicle according to an embodiment of the present invention;

[0032] Figure 8 This is a schematic diagram of a state switching system for an energy storage device in a vehicle according to an embodiment of the present invention. Detailed Implementation

[0033] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0034] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0035] According to an embodiment of the present invention, an embodiment of a state switching method for an energy storage device in a vehicle is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0036] Figure 1 This is a flowchart of a state switching method for an energy storage device in a vehicle according to an embodiment of the present invention, such as... Figure 1 As shown, the method may include the following steps:

[0037] Step S102: Obtain the first state information of the energy storage device in the vehicle and the second state information of the environment in which the vehicle is located.

[0038] In the technical solution provided by step S102 of the present invention, the first state information is used to represent the operating state of the energy storage device in the environment at the current moment. The second state information is used to represent the environmental state affecting the operating state.

[0039] Optionally, the energy storage device can be the vehicle's battery system (also known as a battery), used to store and release electrical energy, and is a core component of the vehicle's energy conversion and transmission. Batteries not only power the vehicle but also affect its performance, driving range, and safety. The aforementioned energy storage device contains multiple battery cells connected through complex circuitry and thermal management systems to collectively power the vehicle.

[0040] Optionally, the first state information can be used to represent the real-time operating status of the energy storage device (battery) under current environmental conditions. This first state information can include various key parameters collected by battery state sensors, such as battery temperature, individual cell voltage, total current, state of charge (SOC), state of health (SOH), gas and pressure monitoring, etc. Specifically, the battery temperature can be monitored using a 16-channel temperature sensor, achieving high accuracy (±0.2℃) and a fast sampling frequency (10Hz), reflecting the battery's thermal state promptly. The individual cell voltage monitors the voltage of each individual battery cell, with fluctuations controlled within ±5mV, helping to detect inconsistencies within the battery and serving as an important indicator of battery health. The total current reflects the battery's charging and discharging current, with accuracy controlled within ±1A, used to monitor the battery's load and prevent thermal runaway caused by overcharging or over-discharging. The SOC can represent the percentage of remaining battery capacity. The SOH reflects the degree of battery capacity degradation, used to assess the battery's health status. The above gas and pressure monitoring can provide information on CO / Gas sensors and internal battery pressure sensors capture gas releases and pressure changes before the battery experiences thermal runaway.

[0041] Optionally, the second state information can refer to external environmental and road condition information that affects the operating state of the energy storage device. This second state information may include, but is not limited to, the vehicle's geographical location information provided by the Global Positioning System (GPS) or the BeiDou Navigation Satellite System, which may include slope and altitude. The vehicle navigation system can analyze driving plans, predict future driving conditions (e.g., rapid acceleration, hill climbing), and adjust thermal management strategies in advance to cope with potential high heat loads. It can also include ambient temperature and wind speed. By monitoring the external ambient temperature (-40℃~85℃) and wind speed (0-20m / s), heat exchange efficiency can be assessed, and the cooling and heating strategies of the thermal management system can be adjusted in a timely manner to ensure battery temperature control under various environments.

[0042] In this embodiment, the first state information of the energy storage device in the vehicle and the second state information of the environment in which the vehicle is located can be obtained.

[0043] Optionally, ensure all sensors and information modules are activated and calibrated, including the battery status sensor, safety monitoring unit, and environmental and road condition module. A 16-channel temperature sensor distributed throughout the battery pack acquires real-time temperature data of individual battery cells and the module at a frequency of 10Hz, with an accuracy controlled within ±0.2℃. Individual battery cell voltage (accuracy ±5mV) and total current (accuracy ±1A) are recorded, while the battery's state of charge (SOC) and state of health (SOH) are also acquired. Utilizing CO / Gas sensors detect any potential gas leaks (1 ppm resolution), and internal pressure sensors monitor pressure changes within the battery pack (±1 kPa accuracy). A dual-mode GPS and BeiDou positioning system acquires the vehicle's precise location coordinates and further analyzes the slope and altitude of the area. The vehicle navigation system provides driving routes and estimated arrival times for the current and future period, helping to predict battery demand. Ambient temperature (range -40°C to 85°C) and wind speed (range 0-20 m / s) are measured to assess external heat exchange efficiency and adjust insulation strategies.

[0044] Optionally, the collected first and second state information are integrated in real time in the central processing unit. The information undergoes preliminary cleaning and processing to remove invalid or abnormal data, ensuring the quality of the input information. A data fusion algorithm is then applied to combine battery status with environmental condition information to form a complete description of the operating status.

[0045] Step S104: Based on the first state information and the second state information, predict the state assessment information of the energy storage device at future moments.

[0046] In the technical solution provided by step S104 of the present invention, the future time is later than the current time.

[0047] Optionally, the state assessment information can be a comprehensive evaluation of the health and safety status of the battery energy storage device at a predicted future moment. This state assessment information can be based on real-time collected data on the battery's internal state (first state information) and the external environment's state (second state information). The core function of this state assessment information is to predict potential health degradation and safety risks the battery may face in advance, thereby providing a basis for the proactive intervention of the thermal management system. This ensures that the battery can maintain optimal operating conditions under various driving conditions and prevents extreme events such as thermal runaway.

[0048] For example, the aforementioned state assessment information may include, but is not limited to, temperature field distribution, heating rate, and thermal runaway risk index. The temperature field distribution can be used to represent the temperature distribution of each cell and module of the battery at future times. Through prediction, the battery's temperature gradient can be accurately determined, overheated or underheated areas can be identified, and cooling or heating measures can be taken in advance to avoid battery performance degradation and safety hazards caused by uneven temperature. The heating rate refers to the rate of temperature rise of the battery as a whole or in a localized area per unit time caused by battery power consumption or power output. Predicting the heating rate is particularly important for identifying instantaneous high-heat conditions such as rapid acceleration and hill climbing. The thermal runaway risk index (0-100) can be used as an assessment indicator to quantify the probability of thermal runaway of a battery under specific environmental and operating conditions. The index ranges from 0 to 100, with higher values ​​indicating greater risk. Through real-time monitoring and prediction, the risk of thermal runaway can be assessed and warned in a timely manner, thereby enabling necessary preventive and control measures to be taken.

[0049] In this embodiment, after obtaining the first state information and the second state information of the energy storage device in the vehicle, the state assessment information of the energy storage device at future times can be predicted based on the first state information and the second state information.

[0050] Optionally, all collected first-state information (such as battery temperature, voltage, current, SOC, SOH, and CO / ) can be aggregated. The input data includes gas concentration, internal pressure, and secondary state information (such as ambient temperature, wind speed, slope, altitude, and driving plan). Quality checks are performed on the input data to remove outliers and missing values, such as temperature fluctuations exceeding reasonable ranges or distorted gas sensor signals. A time-series feature matrix is ​​constructed, and the time-series data is organized using a sliding window technique to ensure model input format compatibility.

[0051] Optionally, a pre-trained Long Short-Term Memory (LSTM) prediction model is loaded. This model can process time-series data and learn the trends of historical battery states and environmental changes. The first and second state information, including multi-dimensional features such as battery state parameters and environmental data, are input into the LSTM model in a pre-processed format. Running the model outputs key state assessment information for the energy storage device over the next few seconds to minutes (depending on the model design), such as temperature field distribution, heating rate, and thermal runaway risk index.

[0052] Optionally, the optimization objectives are defined as minimizing battery temperature fluctuations (within ±2℃), minimizing system energy consumption, and maintaining the thermal runaway risk index within a safe range (e.g., <30). The historical best-practice strategy library is retrieved to find the control strategy that best matches the current prediction results.

[0053] Optionally, the Deep Q-Network (DQN) model generates adjustment commands for cooling, heating, and safety levels based on the prediction results and optimization objectives. The optimal control commands output by the DQN are simulated and executed, predicting changes in the battery state after execution. The effectiveness and accuracy of the control strategy are evaluated by comparing the simulation results with the future state assessment information predicted by the LSTM. After actually executing the control strategy generated by the DQN, the latest data on the battery state is continuously collected. The deviation between the actual state and the predicted state is monitored, and if the deviation exceeds a set threshold (e.g., temperature deviation > 1°C), the model is immediately re-evaluated. The LSTM and DQN model parameters are updated periodically (e.g., every 10 minutes) using gradient descent or similar algorithms to adapt to changes in the environment and battery state, improving the accuracy of long-term predictions.

[0054] In this embodiment, by fusing first and second state information, an LSTM model is used to predict the future state of the energy storage device, and a DQN model is used to optimize decisions and generate control strategies for future moments. The entire process, through real-time data acquisition, model prediction, control strategy formulation, simulation verification, and feedback learning, forms a dynamically optimized closed loop. This ensures that the thermal management system can proactively and accurately respond to various potential future state changes, effectively maintaining the health and safety of the battery. It also demonstrates the significant potential of high-precision prediction and autonomous learning capabilities in improving battery thermal management efficiency.

[0055] Step S106: In response to the status assessment information indicating that the operating state is abnormal, the initial control strategy of the vehicle is adjusted using the performance adjustment factors of the energy storage device to obtain the target control strategy.

[0056] In the technical solution of step S106 of the present invention, the performance adjustment factor is used to adjust the operating performance of the energy storage device. The operating performance of the energy storage device controlled according to the target control strategy is greater than the operating performance of the energy storage device controlled according to the initial control strategy. The target control strategy is used to represent the rules for controlling the vehicle to switch to abnormal states.

[0057] Optionally, an abnormal state can refer to a condition that exceeds the normal operating range during the operation of an energy storage device (such as a battery), which may negatively affect the performance, safety, or lifespan of the energy storage device. In the embodiments of this application, an abnormal state can refer to problems that may be encountered in battery thermal management, such as thermal runaway, i.e., a dangerous state in which the battery temperature rises rapidly to an uncontrollable level, which may cause a fire or explosion. In an abnormal state, key parameters of the battery such as temperature, voltage, gas concentration, and internal pressure may exceed safety thresholds or normal operating ranges.

[0058] Optionally, performance adjustment factors can refer to parameters that directly affect the operating performance of energy storage devices. These parameters are typically related to cooling efficiency, heating efficiency, and the response speed of safety measures in thermal management systems. In the embodiments of this application, performance adjustment factors may involve temperature fluctuations, thermal runaway risk index, system energy consumption, etc. These performance adjustment factors can be dynamically adjusted to cope with abnormal states and ensure that the battery state returns to a safe range.

[0059] Optionally, the initial control strategy described above can be a standard control scheme adopted based on the current operating conditions and environmental conditions when no abnormal state occurs. This initial control strategy may include setpoints for maintaining the normal operating temperature of the battery, stabilizing the state of charge (SOC), and ensuring appropriate cooling and heating. The initial control strategy may trigger cooling or heating based on preset thresholds without considering more complex environmental changes or battery state predictions.

[0060] Optionally, the target control strategy can be an optimized control scheme dynamically generated based on performance adjustment factors when an abnormal state of the energy storage device is detected, to ensure that the energy storage device quickly and safely returns to normal operation. In this embodiment, the target control strategy can be based on the multi-objective optimization principle of "temperature fluctuation ≤ ±2℃, thermal runaway risk index < 30, and minimum system energy consumption," and calculate the optimal control command through an artificial intelligence (AI) decision layer (e.g., a DQN model), such as adjusting the cooling system flow, heating system power, or activating additional safety measures. The target control strategy not only considers immediate battery state adjustments but also future predicted states to improve the overall system response efficiency and safety, and avoid possible future abnormal states.

[0061] In this embodiment, after predicting the state evaluation information based on the first state information and the second state information, if the state evaluation information indicates that the operating state is abnormal, the initial control strategy of the vehicle can be adjusted using the performance adjustment factors of the storage device to obtain the target control strategy.

[0062] Optionally, state assessment information from the perception and decision layers is parsed. This information includes a comprehensive assessment of the current and future battery state, such as temperature field distribution, heating rate, and thermal runaway risk index. The state assessment information is checked to determine if the battery is in an abnormal state. If the thermal runaway risk index exceeds a preset threshold (e.g., reaches or exceeds 30), the battery state is confirmed to have deviated from normal and entered an abnormal state. Once an abnormal state is detected, the system immediately activates the performance adjustment mechanism. Performance adjustment factors involve multiple dimensions such as cooling, heating, and safety protection, specifically including parameters such as cooling flow rate, heating power, and safety level. These parameters are key control points that can be immediately adjusted to cope with the current abnormal state.

[0063] Optionally, the system compares the operational performance provided by a dynamically generated target control strategy based on performance adjustment factors under the current abnormal state with that provided by a static initial control strategy. The target control strategy is formulated based on the principles of ensuring stable battery operation (temperature fluctuation ≤ ±2℃), preventing thermal runaway (risk index < 30), and reducing system energy consumption. Therefore, its operational performance is expected to be superior to the initial control strategy, especially in its ability to correct abnormal states. Based on performance adjustment factors, the system calculates and generates the target control strategy. This target control strategy may include, but is not limited to: increasing cooling flow to rapidly reduce battery temperature, reducing heating power to avoid overheating, or increasing the safety level to enhance monitoring and emergency response. The core of the target control strategy is that it is a dynamically optimized scheme capable of rapidly responding to real-time monitored battery status and environmental conditions.

[0064] In this embodiment, the method described above is a process of dynamically adjusting the performance adjustment factors of the energy storage device after detecting an abnormal state, in order to optimize the control strategy and improve operational performance. This method ensures that the thermal management system can react quickly and effectively control the battery state in the face of abnormal situations such as thermal runaway, preventing further deterioration while minimizing the impact on the overall system energy consumption.

[0065] Step S108: At a future time, in accordance with the target control strategy, control the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state.

[0066] In the technical solution of step S108 of the present invention, the safe state can refer to a state in which the battery operates within its designed performance parameter range, has no risk of thermal runaway, has a stable battery health condition, and does not pose a safety threat to vehicle occupants or the surrounding environment. The safe state may include: stable temperature, low risk of thermal runaway, and effective control of system energy consumption, etc.

[0067] In this embodiment, after adjusting the initial control strategy using the performance adjustment factors of the energy storage device to obtain the target control strategy, the vehicle can be controlled to switch the operating state of the energy storage device from an abnormal state to a safe state at a future time according to the target control strategy.

[0068] Optionally, the AI ​​decision layer (LSTM prediction module and DQN optimization unit) continuously predicts future operating conditions. Based on existing historical data and the road conditions the vehicle will face (such as uphill driving, high-speed driving, and congestion), combined with ambient temperature and other external conditions (such as charging demand), it generates predictive information about future operating conditions. Once it predicts that a future moment may cause the energy storage device to enter an abnormal state (such as an increased risk index of thermal runaway), the system will quickly match or generate a target control strategy based on performance adjustment factors (cooling flow, heating power, safety level, etc.). The target control strategy aims to prevent or mitigate abnormal states and ensure that the battery returns to a safe state.

[0069] Optionally, in the future, the operating parameters of the energy storage device will be adjusted in real time according to the target control strategy: cooling flow adjustment – ​​if the battery temperature is predicted to rise too quickly, the system will instruct the execution layer to increase the flow of the cooling system to accelerate heat dissipation; heating power adjustment – ​​if the expected ambient temperature is extremely low, the heating power can be increased, or the waste heat recovery loop can be activated to ensure that the battery temperature does not drop too low; safety level adjustment – ​​if the predicted thermal runaway risk index is high, the safety protection level can be increased, the explosion-proof valve can be pre-opened, and the high-voltage circuit can be cut off in preparation for unforeseen circumstances. Simultaneously, the status of the energy storage device will be continuously monitored to ensure the effective execution of the target control strategy, and key indicators such as battery temperature, gas concentration, and internal pressure will be measured in real time.

[0070] Optionally, after executing the target control strategy for a period of time, actual effect data will be collected through the feedback layer and compared with the expected target. If it is found that the actual state fails to meet the requirements of the safe state (e.g., battery temperature fluctuation range exceeds ±2℃, thermal runaway risk index is higher than the preset threshold), the system will automatically update the parameters of the LSTM prediction model and DQN optimization unit through gradient descent or other learning algorithms, further adjusting the target control strategy to better fit the real operating conditions. The state of the energy storage device will be reassessed to confirm whether the above state has switched from an abnormal state to a safe state. If the battery state has stabilized, temperature fluctuations are controlled within a reasonable range, the thermal runaway risk index has decreased to a safe level, and system energy consumption is within expectations, then the operating state of the energy storage device can be considered to have successfully transitioned to a safe state. If the state assessment shows that a safe state has not yet been reached, the above steps will continue to be executed, forming a closed-loop feedback until the operating state of the energy storage device fully recovers to normal.

[0071] In this embodiment, through the aforementioned dynamic and intelligent control process, AI technology is used to accurately predict future operating conditions. Combined with real-time data acquisition and analysis, the operating parameters of the energy storage device are automatically adjusted to prevent and correct abnormal states, ensuring that the battery is always in a safe and efficient operating state. The entire process emphasizes the importance of a closed-loop feedback mechanism, achieving a smooth transition from abnormal to safe states, and improving the overall intelligence level and safety assurance capabilities of the thermal management system.

[0072] Steps S102 to S108 of this application, by fusing the first state information of the battery with the second state information of the vehicle environment, utilize advanced predictive algorithms to assess the future state of the energy storage device in advance, significantly improving the accuracy of state switching. Specifically, the embodiments of this application introduce real-time monitoring and analysis based on multi-source data, not limited to temperature and voltage, but also covering early thermal runaway indicators such as gas concentration and internal pressure, enabling early warning before the energy storage device enters an abnormal state. By predicting future state assessment information, the performance adjustment factors of the energy storage device, such as dynamically adjusting heating power, cooling efficiency, and safety level, can be proactively utilized before an abnormal state occurs to optimize the initial control strategy and form a target control strategy. The execution of the above-mentioned target control strategy ensures that the energy storage device can be switched from an abnormal state to a safe state in a timely manner in the future, avoiding thermal management failure caused by state judgment delays, thereby effectively solving the problem of low state switching accuracy in related technologies, improving the efficiency of battery thermal management and the overall safety of the vehicle. This solves the technical problem of low state switching accuracy of energy storage devices in vehicles and achieves the technical effect of improving the state switching accuracy of energy storage devices in vehicles.

[0073] The method described in this embodiment will be further described below.

[0074] As an optional embodiment, step S104, predicting the state assessment information of the energy storage device at a future time based on the first state information and the second state information, includes: performing noise reduction processing on the first state information and the second state information to obtain the noise-reduced first state information and the noise-reduced second state information, wherein the accuracy of the noise-reduced first state information is greater than the accuracy of the first state information before noise reduction, and the accuracy of the noise-reduced second state information is greater than the accuracy of the second state information before noise reduction; constructing a time-series feature matrix based on the noise-reduced first state information and the noise-reduced second state information according to the time window of the target period; and determining the state assessment information using the time-series feature matrix.

[0075] In this embodiment, noise reduction processing can be used to remove random noise and interference signals from the data collected by the sensors, thereby improving the accuracy and reliability of the data. In a battery thermal management system, sensors (such as temperature sensors and gas sensors) may be affected by environmental interference, hardware instability, or electronic noise, resulting in the raw data collected containing errors or irrelevant information. Wavelet transform noise reduction processing can be applied to the first and second state information. By decomposing the signal into wavelet coefficients of different frequencies, coefficients that do not represent meaningful information are identified and filtered out, thereby preserving and enhancing the true components of the signal. The noise-reduced state information (first state information and second state information) is more accurate and reliable.

[0076] Optionally, the target period can refer to the length of the time window used for prediction and control strategy formulation, which is the basic time unit for the thermal management system to perform state assessment and control decisions. In this embodiment, the target period can be set to 10 seconds, which means that the battery state will be reassessed every 10 seconds to predict the state development within the next 10 seconds. The selection of the target period needs to balance real-time performance and computing resources; too short a period may increase the computational burden, while too long a period may lead to response lag. 10 seconds, as a medium-length time window, can achieve fast response while ensuring prediction accuracy. When constructing the time-series feature matrix, the time window of the target period determines the size and information coverage of the feature matrix, thus directly affecting the accuracy of the state assessment information.

[0077] Optionally, the temporal feature matrix can be a data structure used to represent the states of multiple variables within a continuous time period. The temporal feature matrix aggregates the evolution sequence of the first state information (e.g., battery temperature, internal pressure) and the second state information (e.g., ambient temperature, driving conditions) after noise reduction within a certain time window (e.g., 10 seconds, with a step size of 2 seconds). This data can be organized into a multi-dimensional matrix, where each column represents a specific state parameter and each row represents a time point. Through this temporal feature matrix, the dynamic trend of battery state changes over time can be captured, providing sufficient information for the LSTM prediction model to predict the battery state at future moments. Constructing the temporal feature matrix is ​​a crucial step in ensuring prediction accuracy because it not only reflects the current state but also includes past evolutionary history, which is beneficial for the model to learn the patterns of state changes and potential interactions.

[0078] Optionally, during the process of predicting state assessment information based on the first and second state information, the first and second state information can be denoised to obtain denoised first and second state information. A temporal feature matrix can be constructed based on the denoised first and second state information within a target period time window. The state assessment information is then determined using the temporal feature matrix.

[0079] Optionally, first state information (such as battery temperature, voltage, current, etc.) and second state information (such as ambient temperature, driving conditions, driving plan, etc.) are collected from the perception layer. These first and second state information may contain noise or outliers when unprocessed. The data preprocessing unit uses wavelet transform technology to denoise the collected state information. These steps aim to eliminate outlier data points caused by sensor malfunctions, electronic interference, or extreme environmental conditions, improving the accuracy and usability of the information. The denoised information is used for subsequent predictive analysis to ensure the reliability of the prediction results.

[0080] Optionally, the denoised first and second state information is further filtered and organized to ensure that all data are arranged in chronological order and match the time window of the target period. The preprocessing unit constructs a temporal feature matrix based on the denoised state information. With a target period of 10 seconds and a sampling point every 2 seconds, a cyclic window is formed. Each column in the matrix represents a state parameter, and each row corresponds to data at a specific time point. This not only preserves the temporal series characteristics of the state information but also facilitates time-dependent analysis by the AI ​​model to predict future states.

[0081] Optionally, the denoised time-series feature matrix is ​​used as input to the LSTM prediction model. The above steps leverage the efficient processing capabilities of the LSTM model for time-series data, enabling it to predict future state developments based on historical data and the current state. After analyzing the time-series feature matrix, the LSTM model outputs predicted future state assessment information, including key indicators such as battery temperature field distribution, heating rate, and thermal runaway risk index. This predicted information will be used for further control strategy decisions to ensure the battery remains in a safe state when facing upcoming operating conditions.

[0082] In this embodiment of the application, the above method ensures that the battery thermal management system can accurately predict and effectively respond to possible abnormal states in the future. By adjusting the control strategy in real time, the battery is kept in a safe and efficient state, thereby improving the stability and reliability of the entire vehicle system.

[0083] As an optional implementation method, the state assessment information is determined using a time-series feature matrix, including: inputting the time-series feature matrix into the long short-term memory network model corresponding to the energy storage device, and using the long short-term memory network model to output the state assessment information of the future time after the current time for a target duration. The state assessment information includes at least one of the following: distribution information, heating rate, and risk information. The distribution information is used to represent the temperature distribution in the energy storage device, and the risk information is used to represent the risk level of the energy storage device being in an abnormal state.

[0084] In this embodiment, distribution information refers to the temperature distribution within the battery. Since a battery pack consists of multiple battery cells, each cell may generate heat at different rates during charging, discharging, or under different environmental conditions, leading to uneven temperature distribution within the battery pack. Predicting distribution information helps the system identify hot or cold spots within the battery pack in advance, preventing localized overheating or undercooling. For example, the distribution information output by the LSTM prediction model within the next 30 seconds will include the predicted temperature value for each battery cell or module. Based on this distribution information, the coolant flow rate and temperature can be adjusted in advance to ensure a balanced temperature within the battery pack, maintaining it within a safe operating range.

[0085] Optionally, the heating rate can refer to the rate at which the battery generates heat per unit time. A rapid heating rate can indicate that the battery is undergoing a high-load charge / discharge process or is affected by extreme ambient temperatures. Predicting the heating rate can be used to determine whether the battery is about to enter an abnormal state, thereby allowing for timely adjustments to heating or cooling strategies. For example, if the prediction shows that the battery's heating rate will increase significantly in the next 30 seconds, the power of the cooling system can be increased in advance, or the battery's charge / discharge current can be limited to prevent the battery temperature from rising too quickly and reduce the risk of thermal runaway.

[0086] Optionally, risk information can refer to the degree of risk of the battery being in an abnormal state, especially thermal runaway. Risk information can be quantified as an index (e.g., a thermal runaway risk index), with values ​​ranging from 0 to 100; the higher the value, the greater the risk of thermal runaway. By predicting risk information, measures can be taken in advance, such as triggering early warning systems, adjusting battery charging and discharging strategies, or activating safety protection mechanisms to reduce or avoid thermal runaway. Specifically, if the predicted risk information value exceeds a preset safety threshold within the next 30-second time window, the thermal management system will immediately initiate an emergency response procedure to ensure the battery condition quickly returns to a safe level.

[0087] Optionally, the target duration can refer to the time range for state assessment and prediction. In this embodiment, the target duration is set to 30 seconds, which can be based on a deep understanding of the battery's thermal characteristics and consideration of the need for advance warning. A prediction duration of 30 seconds provides the system with sufficient time to respond and adjust to cope with instantaneous heating conditions such as rapid acceleration and hill climbing, while ensuring the accuracy and real-time nature of the prediction. Choosing an appropriate target duration is crucial for the design of a thermal management system; too long a duration may lead to lag, while too short a duration may increase the uncertainty of the prediction. A target duration of 30 seconds balances the accuracy of the prediction with the speed of the system response, providing a reasonable time window for early warning and proactive thermal management.

[0088] Optionally, in the process of determining state assessment information using the time-series feature matrix, the time-series feature matrix can be input into the long short-term memory network model corresponding to the energy storage device, and the long short-term memory network model can be used to output the state assessment information of the future time after the target time after the current time.

[0089] Optionally, the matrix received from the data preprocessing unit contains the evolution sequence of denoised first-state information (e.g., battery temperature, gas concentration, internal pressure) and second-state information (e.g., ambient temperature, driving plan) within a target period (e.g., 10s, step size 2s). A Long Short-Term Memory (LSTM) network model corresponding to the energy storage device is loaded; this model has been trained to predict the future state of the battery. The temporal feature matrix is ​​input into the LSTM model, where each column represents a state parameter and each row represents data at a given time point. The LSTM model begins processing the input temporal feature matrix, utilizing its internal memory units to capture and learn the complex patterns of battery state evolution over time. The model outputs future state assessment information for a target duration (e.g., 30s) after the current moment. The above steps involve multi-step prediction; that is, the model not only predicts the state at the next moment but also predicts for multiple consecutive time points until the target duration ends. State assessment information is extracted from the model's output, including but not limited to distribution information (battery temperature distribution), heating rate (the rate at which the battery heats up), and risk information (the degree of risk of the battery being in an abnormal state). Distribution information indicates the temperature distribution at various points or regions inside the battery. The heating rate quantifies the intensity and trend of battery heating. Risk information uses a quantitative index (such as the thermal runaway risk index, ranging from 0 to 100) to indicate the likelihood of the battery entering abnormal states such as thermal runaway.

[0090] Optionally, the predicted state assessment information, such as distribution information, heating rate, and risk information, is passed to the DQN optimization unit as a basis for decision-making. Based on the predicted state assessment information, the DQN optimization unit generates a control strategy for the next 30 seconds to maintain the battery's optimal operating state, ensuring stable temperature, controlled heating rate, and minimized risk of thermal runaway.

[0091] In this embodiment, the method described above demonstrates how to use a time-series feature matrix and an LSTM network model to predict the state assessment information of an energy storage device for the next 30 seconds, including key indicators such as distribution information, heating rate, and risk information. Through this prediction mechanism, measures can be taken in advance before the battery state changes to optimize temperature control, adjust heating and cooling strategies, and improve safety protection levels, thereby effectively preventing the battery from entering an abnormal state and ensuring the stable operation of the vehicle and the safety of the occupants.

[0092] As an optional embodiment, the performance adjustment factors include at least one of the following: temperature adjustment factor, risk adjustment factor, and energy consumption adjustment factor. The temperature adjustment factor is used to adjust the temperature fluctuation of the energy storage device, the risk adjustment factor is used to adjust the risk level of the energy storage device in an abnormal state, and the energy consumption adjustment information is used to adjust the energy consumption of the energy storage device. In step S106, in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy of the vehicle is adjusted using the performance adjustment factors of the energy storage device to obtain a target control strategy, including at least one of the following: in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy is adjusted using the deep learning model corresponding to the energy storage device with the goal of the temperature adjustment factor being less than or equal to the temperature fluctuation threshold, to obtain a target control strategy; in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy is adjusted using the deep learning model with the goal of the risk adjustment factor being less than the risk level threshold, to obtain a target control strategy; in response to the state assessment information indicating that the operating state is abnormal, the initial control strategy is adjusted using the deep learning model with the goal of minimizing the energy consumption adjustment factor, to obtain a target control strategy.

[0093] In this embodiment, temperature adjustment factors can refer to all controllable variables in the thermal management system used to adjust the degree of battery temperature fluctuation. Risk adjustment factors can refer to factors that can affect the degree of risk of abnormal battery states, mainly including adjustable safety protection measures. Energy consumption adjustment factors can be used to represent the energy consumption control during the operation of the thermal management system. Energy consumption adjustment factors can include the energy efficiency of the cooling system, the power distribution strategy of the heating system, and the optimized control of energy consumption under different battery states and driving conditions.

[0094] Optionally, the deep learning model can be a deep Q-network. The temperature fluctuation threshold can be set to ±2℃, meaning that battery temperature fluctuations should be controlled within this range to maintain battery performance stability and safety. The risk level threshold can be a reference value for the thermal runaway risk index; setting it to <30 indicates that when the thermal runaway risk index is below 30, the battery is in a relatively safe state.

[0095] Optionally, during the process of adjusting the initial control strategy using performance adjustment factors, if the state assessment information indicates an abnormal operating state, a deep learning model can be used to adjust the initial control strategy with the goal of the temperature adjustment factor being less than or equal to the temperature fluctuation threshold, thus obtaining the target control strategy. If the state assessment information indicates an abnormal operating state, a deep learning model can be used to adjust the initial control strategy with the goal of the risk adjustment factor being less than the risk level threshold, thus obtaining the target control strategy. If the state assessment information indicates an abnormal operating state, a deep learning model can be used to adjust the initial control strategy with the goal of minimizing the energy consumption adjustment factor, thus obtaining the target control strategy.

[0096] Optionally, the system receives state assessment information output from the LSTM model, including distribution information, heating rate, and risk information. It determines whether the state assessment information indicates an abnormal operating state for the energy storage device. If the risk information exceeds a preset threshold (e.g., a risk level threshold), or the temperature fluctuation exceeds a predetermined range (e.g., a temperature fluctuation threshold), the device is determined to be in an abnormal state. Based on the specific information of the abnormal state, it identifies the performance adjustment factors that need to be adjusted, namely temperature adjustment factors, risk adjustment factors, or energy consumption adjustment factors. According to the current state of the energy storage device and the predicted future state, it customizes specific adjustment schemes for each performance adjustment factor to address the specific abnormal state.

[0097] Optionally, if the goal is to control temperature fluctuations, a deep learning model (such as DQN) corresponding to the energy storage device is used to adjust the parameters of the cooling or heating system, aiming to ensure that the temperature adjustment factor is less than or equal to the temperature fluctuation threshold (e.g., ±2℃), thus ensuring that the battery temperature remains stable within a safe range. If the goal is to reduce the risk of abnormal states, a deep learning model is also used, setting the risk adjustment factor to be less than the risk level threshold (e.g., <30) as the optimization goal, and adjusting the strategies of safety protection measures, such as the pre-opening degree of the explosion-proof valve and the disconnection speed of the high-voltage circuit, to reduce the risk of thermal runaway. If the goal is to minimize energy consumption, the deep learning model will adjust the initial control strategy with the goal of minimizing the energy consumption adjustment factor, such as optimizing the operating efficiency of the cooling and heating systems to reduce unnecessary energy consumption, while ensuring that temperature fluctuations and risk levels are within a controllable range.

[0098] Optionally, based on the aforementioned target control strategy, control commands are sent to the execution layer of the thermal management system (cooling system, heating system, safety devices) for real-time adjustments. After execution, the battery status is continuously monitored, feedback data is collected to evaluate the effectiveness of the control strategy, and necessary corrections are made based on the feedback to form a closed-loop control process.

[0099] As an optional embodiment, the vehicle includes a cooling system and a heating system. The control strategy includes a cooling strategy and a heating strategy. The cooling strategy represents the rules for performing a cooling operation on the energy storage device, and the heating strategy represents the rules for performing a heating operation on the energy storage device to adjust the abnormal state. Step S108, at a future time, according to the target control strategy, the vehicle is controlled to switch the operating state of the energy storage device from an abnormal state to a safe state, including: at a future time, according to the cooling strategy, controlling the cooling system to perform a cooling operation on the energy storage device to switch the operating state from an abnormal state to a safe state; at a future time, according to the heating strategy, controlling the heating system to perform a heating operation on the energy storage device to switch the operating state from an abnormal state to a safe state.

[0100] In this embodiment, during the switching of the operating state of the control target strategy at a future time, the operating state can be switched by controlling the cooling system to perform a cooling operation on the energy storage device according to the cooling strategy, or by controlling the heating system to perform a heating operation on the energy storage device according to the heating strategy. The cooling system can be a liquid cooling system.

[0101] Optionally, the target control strategy output by the DQN optimization unit is received by the thermal management system. This strategy may include cooling and / or heating strategies, depending on the energy storage device's condition assessment information and target adjustment factors. The system determines whether the current operating state of the energy storage device is abnormal and further analyzes it in conjunction with future predicted condition assessment information (such as temperature and risk index). If the abnormal state is caused by excessively high temperature, a cooling strategy is selected; if it is caused by excessively low temperature, a heating strategy is selected; if both temperature anomalies and energy consumption issues exist simultaneously, a comprehensive consideration of cooling and heating strategies, as well as energy consumption adjustment factors, may be necessary.

[0102] Optionally, the cooling strategy within the target control strategy can be analyzed. This may include precise control requirements for battery temperature, such as controlling temperature fluctuations within ±2°C or lower. Based on the cooling strategy, the thermal management system issues commands to the cooling system, adjusting coolant flow, electronic expansion valve opening, etc., to rapidly reduce battery temperature. For example, in scenarios where rapid acceleration or hill climbing causes a rapid rise in battery temperature, the system may need to increase the coolant flow to its maximum value to quickly cool down. The heating strategy within the target control strategy can also be analyzed. This may include commands to minimize energy consumption while heating the battery to a suitable operating temperature. Based on the heating strategy, the thermal management system issues commands to the heating system, such as activating the PTC heater, adjusting its power, or utilizing waste heat from the motor for temperature increase. For example, in low-temperature environments, the system may need to use maximum power PTC heating or fully utilize waste heat from the motor while controlling the fan speed to quickly raise the battery temperature to its optimal operating range.

[0103] Optionally, by adjusting the cooling or heating system, the operating state of the energy storage device can transition from an abnormal state to a safe state. After performing cooling or heating operations, the battery status is continuously monitored, and feedback data, such as temperature, gas concentration, and internal pressure, are collected to evaluate the effectiveness of the control strategy. If the battery temperature does not return to a safe range after operating according to the cooling strategy, or if the temperature becomes too high after heating operations, the system will further correct the strategy through a feedback mechanism until the battery status reaches a safe standard.

[0104] Optionally, the deviation between the actual state and the target state, such as the difference between the temperature and the target temperature, is calculated. If the deviation exceeds a preset range, the system will adjust the parameters of the DQN model to optimize the cooling or heating strategy for future time periods. This process continues to ensure that the system can learn from each control operation and gradually improve the accuracy and efficiency of the strategy.

[0105] In the embodiments of this application, the thermal management system can intelligently and dynamically adjust the operation of the cooling system or heating system according to the real-time status and predicted future status of the energy storage device, so as to ensure that the battery status switches from an abnormal state to a safe state in a timely manner, while minimizing energy consumption, maintaining battery performance and extending service life.

[0106] As an optional embodiment, the vehicle also includes a safety actuator. The control strategy includes a first control strategy and a second control strategy. The first control strategy represents a rule for adjusting the internal pressure of the energy storage device, and the second control strategy represents a rule for disconnecting the power to the energy storage device. Step S108 involves controlling the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state at a future time, according to the target control strategy. This includes: at a future time, according to the first control strategy, controlling the explosion-proof valve of the safety actuator to adjust the internal pressure of the energy storage device to switch the operating state from an abnormal state to a safe state, wherein the adjusted internal pressure of the energy storage device is lower than the internal pressure of the energy storage device before adjustment; and at a future time, according to the second control strategy, controlling the relay of the safety actuator to disconnect the circuit between the energy storage device and the electrical system in the vehicle to switch the operating state from an abnormal state to a safe state.

[0107] In this embodiment, the first control strategy can refer to a series of rules and methods adopted by the system when an abnormal increase in the internal pressure of the energy storage device is detected, to control the explosion-proof valve of the safety actuator, thereby adjusting the internal pressure of the battery pack to a safe level. The second control strategy can refer to the strategy adopted by the system when an electrical safety risk is identified in the energy storage device, used to control the relay of the safety actuator to ensure that the energy storage device is disconnected from the vehicle's electrical system. The safety actuator can be a component in the thermal management system used to protect battery safety in emergency situations; the aforementioned safety actuator can include explosion-proof valves and relays, etc. The aforementioned explosion-proof valve can be an active explosion-proof valve, a safety device designed to automatically or controllably open and release internal pressure when the internal pressure of the battery exceeds a safe threshold, preventing the battery pack from exploding or physically rupturing due to excessive internal pressure. The relay can be a high-voltage relay, an electronic device used to control the on / off state of high-voltage electrical circuits. In the second control strategy of the thermal management system, when the system determines that there is an electrical safety risk to the battery, it can quickly cut off the circuit connection between the battery and the vehicle's electrical system by controlling the state of the relay (open or closed), thereby preventing the current from continuing to flow under abnormal conditions, thus protecting the battery and the vehicle's electrical equipment and preventing accidents from occurring.

[0108] Optionally, the electrical system can refer to the collection of all components and wiring in a vehicle involved in the transmission, storage, and use of electricity, including but not limited to the battery, powertrain, air conditioning system, entertainment system, lighting, and charging system. The second control strategy of the thermal management system focuses on the connection circuits between the battery and the electrical system. By disconnecting these circuits in a timely manner, it can effectively prevent abnormal battery conditions from affecting the entire vehicle's electrical system, thereby improving the vehicle's electrical safety reliability.

[0109] Optionally, during the process of controlling the vehicle to switch operating states according to the target control strategy, at a future time, according to the first control strategy, the explosion-proof valve of the safety actuator can be controlled to adjust the internal pressure of the energy storage device to switch operating states. Alternatively, according to the second control strategy, the relay of the safety actuator can be controlled to disconnect the circuit between the energy storage device and the electrical system in the vehicle to switch operating states.

[0110] Optionally, the thermal management system receives real-time battery status information and predictive data from the sensing and decision-making layers. Based on the real-time and predictive data, it assesses whether the battery's current operating state is abnormal, such as an abnormally high internal pressure or an electrical safety risk. If the battery is determined to be in an abnormal state, the specific type of abnormality is further determined—whether it is related to internal pressure or electrical safety. Based on the current internal pressure and risk assessment, the decision-making layer generates a first control strategy, which includes parameters such as the opening degree and duration of the explosion-proof valve. According to the first control strategy, the system sends a control command to the active explosion-proof valve in the safety actuator to adjust its opening degree to release excess pressure inside the battery pack. After executing the strategy, the internal pressure of the battery pack is continuously monitored to ensure that the adjusted pressure is lower than the pre-adjustment level, reaching below the safety threshold. Data on changes in internal pressure after the explosion-proof valve's action is collected and fed back to the decision-making layer for subsequent analysis and strategy optimization. It is confirmed whether the internal pressure of the battery pack has been adjusted to a safe level. If it has dropped to a safe range, the abnormal state is alleviated; otherwise, the strategy may need to be repeated or other emergency measures may need to be taken.

[0111] Optionally, the decision-maker generates a second control strategy based on an electrical safety risk assessment. This strategy involves operating instructions for the high-voltage relay. According to the second control strategy, the system sends instructions to the high-voltage relay in the safety actuator to disconnect the battery from the vehicle's electrical system. After the control strategy is executed, the system verifies whether the electrical system has been successfully disconnected, such as checking whether the battery voltage has dropped below a safe level to confirm that the battery has switched from an abnormal state to a safe state. Electrical state data after the power-off operation, including battery voltage and current, is collected and fed back to the decision-maker for analysis. It is confirmed whether the battery has been successfully disconnected and is in an electrically safe state. If so, the abnormal state is resolved; if not, a re-inspection may be necessary, and other measures may need to be taken to ensure safety.

[0112] In this embodiment, the thermal management system, through the aforementioned method, can intelligently identify and respond to abnormal battery conditions. By using the explosion-proof valve and high-voltage relay of the safety actuator, it dynamically adjusts the internal pressure and electrical connection status to ensure the battery switches from an abnormal state to a safe state, thereby protecting the safety of the vehicle and its occupants. Furthermore, through real-time status monitoring and feedback mechanisms, the system can continuously optimize its control strategy, improving its adaptability and effectiveness under various complex operating conditions.

[0113] As an optional embodiment, the method further includes: after the energy storage device switches from an abnormal state to a safe state, acquiring first state information of the energy storage device in the safe state; determining the temperature error of the energy storage device based on the first state information; an adjustment step, in response to the temperature error being greater than an error threshold, determining the future time as the current time, using gradient descent to adjust the model parameters of the deep learning model corresponding to the energy storage device at the current time, and / or the model parameters of the long short-term memory network model of the energy storage device at the current time, to obtain an adjustment result; in response to the adjustment result at the current time being that the temperature error is greater than the error threshold, determining the future time of the target duration after the current time, and returning to execute from the beginning of the adjustment step until the adjustment result is that the temperature error is less than the error threshold.

[0114] In this embodiment, after the energy storage device switches from an abnormal state to a safe state, the first state information of the energy storage device in the safe state can be obtained. Based on the first state information, the temperature error of the energy storage device can be determined. If the temperature error is greater than the error threshold, the future time can be determined as the current time, and the model parameters of the deep learning model corresponding to the energy storage device at the current time, or the model parameters of the long short-term memory network model of the energy storage device at the current time, can be adjusted to obtain the adjustment result. If the temperature error corresponding to the adjustment result is greater than the error threshold, the future time of the target duration after the current time can be determined, and the process can be returned from the adjustment step until the temperature error is less than the error threshold.

[0115] Optionally, after the energy storage device switches to a safe state, the thermal management system begins the next stage of monitoring and parameter optimization. It collects real-time status information of the energy storage device in the safe state, such as battery temperature, ambient temperature, and battery voltage. Based on this initial status information, it calculates the error between the current actual temperature and the target temperature, where the target temperature is the expected temperature predicted by a deep learning model for a future time. It then determines whether the current temperature error exceeds a preset error threshold. If the temperature error exceeds the threshold, it indicates a deviation in temperature control, requiring further optimization. If the temperature error is greater than the error threshold, the current time is identified as the point requiring adjustment. Gradient descent is used to adjust the model parameters of the deep learning model (e.g., DQN) and / or the long short-term memory network model (e.g., LSTM) of the energy storage device to reduce temperature error and improve the model's prediction and control accuracy.

[0116] Optionally, the effectiveness of the adjusted model parameters is evaluated, i.e., whether the temperature error at the current moment has decreased to below the error threshold under the control of the adjusted model. If, after adjustment, the temperature error at the current moment is still greater than the error threshold, the system will determine future moments within a target time period after the current moment and enter the next iteration adjustment process. The temperature error is re-evaluated, and if it is still greater than the error threshold, the deep learning model parameters are adjusted again until the temperature error is less than the error threshold, achieving the expected target for model prediction and control accuracy. The above process is repeated until the temperature error meets the condition, i.e., reaches or falls below the preset threshold, thereby ensuring that the model can continuously optimize over time, improving the control performance and accuracy of the thermal management system.

[0117] In this embodiment, through continuous temperature error monitoring and model parameter adjustment, the thermal management system can intelligently optimize the performance of the deep learning model and the long short-term memory network model, ensuring high precision in battery temperature control, thereby improving the battery's performance and safety under different environments.

[0118] The technical solutions of the embodiments of the present invention will be illustrated below with reference to preferred embodiments.

[0119] Currently, most automotive power battery thermal management systems employ "threshold-triggered" control logic. Related technical shortcomings may include: delayed safety warnings, reliance solely on "temperature-voltage" dual parameters, and failure to incorporate battery gas generation (CO, etc.). Early thermal runaway characteristics such as internal pressure are not detected, with a warning lag of ≥20s, making it unable to cope with instantaneous heating conditions such as rapid acceleration and hill climbing; the heat preservation strategy is rigid, using fixed-power PTC heating without dynamic adjustment based on ambient temperature, SOC, and driving plans (such as navigation charging needs), resulting in a range loss rate of >30% at -20℃; poor control coordination, with cooling, heating, and safety modules operating independently without a unified decision-making center, increasing system energy consumption by an additional 15%-20%; lacking predictive and self-learning capabilities, relying solely on real-time parameter feedback control without incorporating road condition prediction and model self-learning, resulting in control accuracy ≤±5℃ and limited battery cycle life (<2000 cycles).

[0120] In this embodiment, a closed-loop system of "prediction-decision-execution-feedback" is constructed through "multi-source data fusion + AI prediction and decision-making + closed-loop feedback correction" to achieve dynamic matching of "road conditions - battery status - thermal management strategy," thus solving the problems of lag and rigidity in existing technologies. The technical problems to be solved are: achieving early warning of thermal runaway (≥50s in advance) to reduce safety risks under extreme operating conditions; dynamically optimizing the insulation strategy to reduce the range loss rate to less than 10% in -20℃ environments; establishing a collaborative control mechanism for cooling, heating, and safety protection to achieve multi-objective optimization of "temperature stability - energy consumption - lifespan"; and improving temperature control accuracy to ±2℃ to extend battery cycle life to over 2500 cycles.

[0121] The embodiments of the present invention will be further described below.

[0122] Figure 2 This is a flowchart of an artificial intelligence-based battery thermal management system and a multimodal main control closed loop according to an embodiment of the present invention, such as... Figure 2 As shown, the data acquisition (node ​​A) synchronously acquires multi-source information at a frequency of 10Hz—battery parameters (temperature, voltage, SOC, etc.) reflect the current state of the battery, road condition data (slope, vehicle speed) predicts the future heating trend under operating conditions, and environmental data (temperature, wind speed) corrects the heat exchange efficiency, providing comprehensive input for subsequent decision-making; Preprocessing (node ​​B): Wavelet transform is used to remove sensor outliers (such as temperature jumps > 2℃ / s) to avoid interfering with the model; the data is then normalized to the [0,1] interval to construct a time-series feature matrix with a 10s time window, adapting to the time-series data input requirements of the LSTM model; LSTM prediction (node ​​C): Based on historical 60s data and future 5km road condition prediction, the core parameters for the next 30s are output—temperature field distribution (accurate to individual battery cells), heating rate (identifying instantaneous high-heat scenarios such as rapid acceleration / climbing), and thermal runaway risk index (0-100, capturing safety in advance). (Hidden dangers), solving the "lag" problem of traditional control; DQN optimization decision (D node): with "temperature stability (±2℃), safety risk <30, and lowest energy consumption" as multiple objectives, it matches instructions from a library of 1000+ historical best strategies, such as cooling system water pump speed, PTC heating power, and safety level, to achieve "multi-objective collaborative optimization"; Actuator (E node): drives liquid cooling, heating, and safety devices according to instructions—liquid cooling system zone control (differentiated heat dissipation for cells / modules), heating system dual-source switching (prioritizing waste heat energy saving), and safety actuator rapid response (explosion-proof valve opens <50ms) to ensure that instructions are implemented; Feedback correction (F node): collects the deviation between the actual temperature and the predicted temperature after execution. If the deviation is >1℃, it updates the LSTM and DQN model parameters through gradient descent (once every 10 minutes), forming a "prediction-execution-correction" closed loop to improve the model's adaptability under different operating conditions.

[0123] Figure 3 This is a schematic diagram of an artificial intelligence-based battery thermal management system and a multimodal safety control sub-strategy according to an embodiment of the present invention, as shown below. Figure 3 As shown, the three-dimensional feature acquisition (S1 node) selects three early thermal runaway features: temperature, gas, and pressure. Temperature features (T_max), the highest single-cell temperature and heating rate (dT / dt), directly reflect the battery's heating intensity; (dT / dt > 2℃ / min) indicates an anomaly. Gas features (CO concentration), CO is released before battery thermal runaway; 1ppm resolution can capture early gas production. Pressure features (pack pressure), electrolyte decomposition and gas production cause a sudden pressure increase; ±1kPa accuracy provides early warning. Risk index calculation (S2 node): A weighted formula (Risk = 0.4 × T_score + 0.3 × Gas_sorce + 0.3 × P_score) is used. Temperature has the highest weight (0.4) because it is the most direct safety indicator. Gas and pressure (0.3 each) supplement early hidden risks and avoid... Single parameter misjudgment; graded judgment and execution (S3-S7 nodes), matching differentiated actions according to risk level, balancing safety and practicality—Risk < 30 (low risk): normal monitoring and data storage, avoiding excessive intervention; 30 ≤ Risk < 60 (medium risk): instrument warning prompts the driver, limits the charge / discharge rate to 0.8C (reduce heat generation), increases cooling power by 10% (active cooling); 60 ≤ Risk < 80 (high risk): enhanced audible and visual alarm, limits the rate to 0.5C, activates the explosion-proof valve preparatory mode (shortening response time), pushes warning to the cloud platform (facilitating remote monitoring); Risk ≥ 80 (emergency risk): cuts off the high-voltage circuit (avoids short circuit), activates maximum flow cooling (suppresses heat diffusion), opens the explosion-proof valve (releases pressure), triggers door unlocking (ensures personnel safety), covering the entire process of emergency response to thermal runaway.

[0124] Figure 4This is a schematic diagram of a heat preservation control sub-strategy according to an embodiment of the present invention, as shown in Figure 4. Trigger condition determination (H1-H2 nodes): Combining the dual dimensions of "temperature + driving plan"—temperature threshold, ambient temperature (T_e < 15℃) or average battery temperature (T_b < 20℃) (the optimal battery operating temperature is 25-35℃; below 20℃, capacity decay is >10%); driving plan, if the navigation shows a charging requirement in the next 10 minutes, the heat preservation start threshold can be appropriately lowered (e.g., T_b < 18℃) to avoid overheating during charging; PTC power calculation (H3 node): Dynamically adjusting power using the formula (P = k × (25 - T_b) × (1 - SOC / 100))—environmental coefficient k: (k = 1.2) when (T_e < -10℃) (lower temperatures require stronger heating), otherwise (k = 1.0); ((25 - T_b)): Matching the base power according to the difference between the battery and the optimal temperature (25℃); ((1 - SOC / 100) OC / 100): When SOC is high (e.g., >80%), power is reduced to avoid overcharging and heating of the battery; Intelligent heat source switching (H4-H6 nodes): Prioritize the use of motor waste heat (H5 node) - when the motor water temperature is >40℃, switch to the waste heat recovery circuit (waste heat generated during motor operation), turn off PTC, and reduce the energy consumption of the whole vehicle; when waste heat is insufficient (e.g., in the early stage of cold start), start PTC (H6 node) to ensure heating efficiency; Passive heat preservation assistance (H7 node): Adjust the speed of the circulating fan in the battery pack (0-1500rpm) to maintain the air temperature gradient of the insulation layer <5℃, reduce the heat exchange loss between the battery and the outside world, and assist in active heating to reduce energy consumption; Closed-loop stop condition (H8-H9 nodes): Dual stop logic, when the temperature reaches the standard: (T_b≥25℃) and (T_e≥15℃), the battery returns to the best working state; Operating condition switching: when entering fast charging mode (charging current >50A), heat will be generated during the charging process, and the heat preservation is turned off to avoid overheating, while reducing charging energy consumption.

[0125] Figure 5(a) is a schematic diagram of a battery system according to an embodiment of the present invention. As shown in Figure 5(a), it can be a battery system applying the temperature regulation method of the embodiment of the present invention, which may include an upper battery housing 51 and a lower battery housing 52. Figure 5(b) is a schematic diagram of another battery system according to an embodiment of the present invention. As shown in Figure 5(b), it may include a battery high-voltage matching system 53 and a battery management system 54. The battery high-voltage matching system 53 can be used to switch the high-voltage circuit on and off. For example, if an abnormal state of the battery is detected, it can control the disconnection of the battery from the circuit of other electrical systems in the vehicle. The battery management system 54 can be used for operation algorithms and comprehensive control. For example, it can execute the state switching method of the embodiment of the present invention.

[0126] Figure 6(a) is a schematic diagram of the internal structure of a battery system according to an embodiment of the present invention. As shown in Figure 6(a), the internal structure of the battery system may include a battery cooling and heating system 61 and a battery cell (energy storage unit) 62. The battery cooling and heating system 61 can cool or heat the battery based on control commands (target control strategy) generated by the battery management system 54.

[0127] Figure 6(b) is a schematic diagram of the internal structure of another battery system according to an embodiment of the present invention. As shown in Figure 6(b), the battery system may further include a signal acquisition system 63 and a battery safety protection system 64. The signal acquisition system 63 can be used to collect information related to battery temperature regulation, such as the battery's first state information, etc. There are no specific limitations here. The relevant information can be input into the battery management system 54 to determine whether a state switch is required.

[0128] It should be noted that the above-described battery system architecture and internal component structure are merely illustrative examples and are not subject to specific limitations. Any battery system and its internal structure that can apply the state switching method in the embodiments of the present invention are within the protection scope of the embodiments of the present invention.

[0129] In this application embodiment, the safety performance is as follows: thermal runaway warning is advanced to 50s, false alarm rate is <0.1%, high voltage cut-off response time is ≤100ms, meeting the safety requirements of GB38031-2020; thermal insulation efficiency is reduced from 30% to 8% in -20℃ environment, and PTC energy consumption is reduced by 25% (contributed by waste heat recovery); temperature control has a temperature fluctuation range of ±2℃ (traditional technology ±5℃), and battery cycle life is extended to 2500 times (an improvement of 25%); adaptive capability: through closed-loop learning, the control accuracy retention rate is >90% in extreme environments (high altitude, cold regions), adapting to different vehicle models.

[0130] For example, in an extreme low-temperature scenario (Te=-25℃, SOC=30%, urban congestion): Data acquisition: The perception layer collects battery temperature at 10℃, ambient temperature at -25℃, and congestion coefficient at 0.8 (average vehicle speed 15km / h, no charging plan); Preprocessing: Wavelet transform removes one set of abnormal temperature data (jumps of 1.5℃ / s, determined to be sensor interference), generating a standardized feature vector [10,30,-25,0.8,...]; LSTM prediction: Outputs that the battery temperature will drop to 8℃ in the next 30 seconds, with a thermal runaway risk index of 12 (low risk) and a heat preservation requirement coefficient of 0.9; DQN decision: Generates instructions. —PTC power 3.6kW (k=1.2, P=1.2×(25-10)×(1-30 / 100)=3.6kW), circulating fan 800rpm, current limit ≤80A (to avoid high current heat generation fluctuations); feedback is executed, and the battery temperature is fed back to 15℃ after 30s (error 0.3℃<1℃, no need for emergency correction), the model fine-tunes the k value to 1.1, and the PTC power drops to 3.3kW; stable control, when the motor water temperature rises to 42℃, switch to waste heat recovery heating, turn off PTC, fan speed drops to 500rpm, and maintain the battery temperature at 18-22℃ (fluctuation ±2℃).

[0131] According to embodiments of the present invention, a state switching device for an energy storage device in a vehicle is also provided. It should be noted that this state switching device for an energy storage device in a vehicle can be used to execute the state switching method for an energy storage device in a vehicle described in the above embodiments.

[0132] Figure 7 This is a schematic diagram of a state switching device for an energy storage device in a vehicle according to an embodiment of the present invention, such as... Figure 7 As shown, the state switching device 700 of the energy storage device in the vehicle may include: an acquisition unit 702, a prediction unit 704, an adjustment unit 706, and a control unit 708.

[0133] The acquisition unit 702 is used to acquire the first state information of the energy storage device in the vehicle and the second state information of the environment in which the vehicle is located.

[0134] The prediction unit 704 is used to predict the state assessment information of the energy storage device at a future time based on the first state information and the second state information.

[0135] The adjustment unit 706 is used to adjust the initial control strategy of the vehicle in response to the state assessment information indicating that the energy storage device is in an abnormal state, by utilizing the performance adjustment factors of the energy storage device, to obtain the target control strategy.

[0136] The control unit 708 is used to control the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state at a future time, according to the target control strategy.

[0137] In this embodiment of the invention, the acquisition unit 702 acquires first state information of the energy storage device in the vehicle and second state information of the vehicle's environment. The prediction unit 704 predicts the state assessment information of the energy storage device at a future time based on the first and second state information. The adjustment unit 706, responding to the state assessment information indicating that the energy storage device is in an abnormal state, adjusts the vehicle's initial control strategy using the performance adjustment factors of the energy storage device to obtain a target control strategy. The control unit 708, at a future time, controls the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state according to the target control strategy. This solves the technical problem of low accuracy in state switching of energy storage devices in vehicles and achieves the technical effect of improving the accuracy of state switching of energy storage devices in vehicles.

[0138] Figure 8 This is a schematic diagram of a state switching system for an energy storage device in a vehicle according to an embodiment of the present invention, such as... Figure 8 As shown, the state switching system 800 of the energy storage device in the vehicle may include: a perception layer 802, a decision layer 804, and an execution layer 806.

[0139] The perception layer 802 is used to obtain the first state information of the energy storage device in the vehicle and the second state information of the environment in which the vehicle is located.

[0140] The decision layer 804 is used to predict the state assessment information of the energy storage device at future times based on the first state information and the second state information; in response to the state assessment information indicating that the energy storage device is in an abnormal state, the initial control strategy of the vehicle is adjusted using the performance adjustment factors of the energy storage device to obtain the target control strategy.

[0141] The execution layer 806 is used to control the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state in the future, according to the target control strategy.

[0142] In this embodiment of the invention, the perception layer 802 acquires first state information of the energy storage device in the vehicle and second state information of the vehicle's environment. Based on the first and second state information, the decision layer 804 predicts the future state assessment information of the energy storage device. In response to the state assessment information indicating that the energy storage device is in an abnormal state, the initial control strategy of the vehicle is adjusted using the performance adjustment factors of the energy storage device to obtain a target control strategy. At a future time, the execution layer 806 controls the vehicle to switch the operating state of the energy storage device from an abnormal state to a safe state according to the target control strategy. This solves the technical problem of low accuracy in state switching of energy storage devices in vehicles and achieves the technical effect of improving the accuracy of state switching of energy storage devices in vehicles.

[0143] The method described in this embodiment will be further described below.

[0144] As an optional embodiment, the system further includes: a feedback layer, used to acquire first state information of the energy storage device in the safe state after the energy storage device switches from an abnormal state to a safe state; determine the temperature error of the energy storage device based on the first state information; and adjust the model parameters of the deep learning model corresponding to the energy storage device and / or the model parameters of the long short-term memory network model of the energy storage device in response to the temperature error being greater than the error threshold.

[0145] According to embodiments of the present invention, a computer-readable storage medium is also provided, the storage medium including a stored program, wherein the program executes the methods described in the embodiments of the present invention.

[0146] According to an embodiment of the present invention, a processor is also provided for running a program, wherein the program executes the methods described in the embodiments of the present invention during runtime.

[0147] According to another aspect of the present invention, an electronic device is also provided. The electronic device includes a memory and a processor, the memory storing a computer program, and the processor being configured to run the computer program to perform the methods described in the embodiments of the present invention.

[0148] According to another aspect of the present invention, a computer program product is also provided. This computer program product includes a computer program that, when executed by a processor, implements the methods described above in the embodiments of the present invention.

[0149] According to another aspect of the present invention, a vehicle is also provided. The vehicle includes a memory and a processor. The memory stores an executable program; the processor runs the program, which, when executed, implements the methods described in the embodiments of the present invention.

[0150] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0151] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0152] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0153] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0154] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0155] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method of switching the state of an energy storage device in a vehicle, characterized by, The method comprises the following steps: acquiring first state information of a storage energy device in a vehicle and second state information of an environment in which the vehicle is located, wherein the first state information is used to represent an operating state of the storage energy device in the environment at a current time, and the second state information is used to represent an environmental state that affects the operating state; based on the first state information and the second state information, predicting state evaluation information of the storage energy device at a future time, wherein the future time is later than the current time; in response to the state evaluation information indicating that the operating state is an abnormal state, adjusting an initial control strategy of the vehicle by using a performance adjustment factor of the storage energy device to obtain a target control strategy, wherein the performance adjustment factor is used to adjust the operating performance of the storage energy device, and the operating performance of the storage energy device controlled according to the target control strategy is greater than the operating performance of the storage energy device controlled according to the initial control strategy, and the target control strategy is used to represent a rule for controlling the vehicle to switch from the abnormal state to a safe state; at the future time, controlling the vehicle to switch the operating state of the storage energy device from the abnormal state to the safe state according to the target control strategy.

2. The method of claim 1, wherein, Based on the first state information and the second state information, predicting state evaluation information of the storage energy device at a future time, comprises: performing noise reduction processing on the first state information and the second state information to obtain noise-reduced first state information and noise-reduced second state information, wherein the accuracy of the noise-reduced first state information is greater than the accuracy of the first state information before noise reduction, and the accuracy of the noise-reduced second state information is greater than the accuracy of the second state information before noise reduction; based on the noise-reduced first state information and the noise-reduced second state information, constructing a time sequence feature matrix according to a time window of a target period; determining the state evaluation information by using the time sequence feature matrix.

3. The method of claim 2, wherein, Determining the state evaluation information by using the time sequence feature matrix comprises: inputting the time sequence feature matrix into a long short-term memory network model corresponding to the storage energy device, and outputting the state evaluation information of the future time after a target time length from the current time by using the long short-term memory network model, wherein the state evaluation information comprises at least one of the following: distribution information, heat generation rate and risk information, the distribution information is used to represent the distribution of temperature in the storage energy device, and the risk information is used to represent the risk degree of the storage energy device in the abnormal state.

4. The method of claim 1, wherein, The performance adjustment factor includes at least one of a temperature adjustment factor, a risk adjustment factor, and an energy consumption adjustment factor, the temperature adjustment factor is used to adjust the temperature fluctuation degree of the energy storage device, the risk adjustment factor is used to adjust the risk degree of the energy storage device in the abnormal state, and the energy consumption adjustment information is used to adjust the energy consumption of the energy storage device, in response to the state evaluation information being the running state being the abnormal state, adjusting the initial control strategy of the vehicle by using the performance adjustment factor of the energy storage device to obtain a target control strategy, including at least one of: In response to the state evaluation information being the running state being the abnormal state, adjusting the initial control strategy by using the corresponding deep learning model of the energy storage device to obtain the target control strategy, with the temperature adjustment factor being less than or equal to a temperature fluctuation threshold; In response to the state evaluation information being the running state being the abnormal state, adjusting the initial control strategy by using the deep learning model to obtain the target control strategy, with the risk adjustment factor being less than a risk degree threshold; In response to the state evaluation information being the running state being the abnormal state, adjusting the initial control strategy by using the deep learning model to obtain the target control strategy, with the energy consumption adjustment factor being minimized.

5. The method of claim 1, wherein, The vehicle includes a cooling system and a heating system, the control strategy includes a cooling strategy and a heating strategy, the cooling strategy is used to represent the rules of performing a cooling operation on the energy storage device, and the heating strategy is used to represent the rules of performing a heating operation on the energy storage device to adjust the abnormal state, in the future moment, the running state of the energy storage device is switched from the abnormal state to a safe state according to the target control strategy, including: In the future moment, the cooling system performs the cooling operation on the energy storage device according to the cooling strategy, so as to switch the running state from the abnormal state to the safe state; In the future moment, the heating system performs the heating operation on the energy storage device according to the heating strategy, so as to switch the running state from the abnormal state to the safe state.

6. The method of claim 5, wherein, The vehicle further includes a safety actuator, the control strategy includes a first control strategy and a second control strategy, the first control strategy is used to represent the rules of adjusting the internal pressure of the energy storage device, and the second control strategy is used to represent the rules of powering off the energy storage device, in the future moment, the running state of the energy storage device is switched from the abnormal state to a safe state according to the target control strategy, including: In the future moment, the safety actuator adjusts the internal pressure of the energy storage device by controlling the explosion-proof valve according to the first control strategy, so as to switch the running state from the abnormal state to the safe state, wherein the adjusted internal pressure of the energy storage device is less than the internal pressure of the energy storage device before adjustment; At the future moment, according to the second control strategy, a relay of the safety actuator is controlled to disconnect the energy storage device from an electrical circuit of an electrical system in the vehicle to switch the operating state from the abnormal state to the safe state.

7. The method according to any one of claims 1 to 6, characterized in that, The method further comprises: After the energy storage device is switched from the abnormal state to the safe state, the first state information of the energy storage device in the safe state is acquired; Based on the first state information, a temperature error of the energy storage device is determined; In response to the temperature error being greater than an error threshold, the future moment is determined as the current moment, and a gradient descent method is used to adjust model parameters of a deep learning model corresponding to the energy storage device at the current moment and / or model parameters of a long short-term memory network model of the energy storage device at the current moment to obtain an adjustment result; In response to the adjustment result at the current moment being that the temperature error is greater than the error threshold, the future moment of a target time length after the current moment is determined, and the adjustment step is returned to be executed until the adjustment result is that the temperature error is less than the error threshold.

8. A state switching device for an energy storage device in a vehicle, characterized in that, The device comprises: An acquisition unit is configured to acquire first state information of an energy storage device in a vehicle and second state information of an environment in which the vehicle is located, wherein the first state information is used to indicate an operating state of the energy storage device in the environment at a current moment, and the second state information is used to indicate an environmental state that affects the operating state; A prediction unit is configured to predict state evaluation information of the energy storage device at a future moment based on the first state information and the second state information, wherein the future moment is later than the current moment; An adjustment unit is configured to, in response to the state evaluation information being that the energy storage device is in an abnormal state, adjust an initial control strategy of the vehicle by using a performance adjustment factor of the energy storage device to obtain a target control strategy, wherein the performance adjustment factor is used to adjust an operating performance of the energy storage device, the operating performance of the energy storage device controlled according to the target control strategy is greater than the operating performance of the energy storage device controlled according to the initial control strategy, and the target control strategy is used to indicate a rule for controlling the vehicle to switch the abnormal state; A control unit is configured to, at the future moment, control the vehicle to switch the operating state of the energy storage device from the abnormal state to a safe state according to the target control strategy.

9. A state switching system for an energy storage device in a vehicle, characterized in that, The system comprises: A perception layer is configured to acquire first state information of an energy storage device in a vehicle and second state information of an environment in which the vehicle is located, wherein the first state information is used to indicate an operating state of the energy storage device in the environment at a current moment, and the second state information is used to indicate an environmental state that affects the operating state; The decision layer is configured to predict state evaluation information of the energy storage device at a future time based on the first state information and the second state information, wherein the future time is later than the current time; in response to the state evaluation information indicating that the energy storage device is in an abnormal state, adjust an initial control strategy of the vehicle by using a performance adjustment factor of the energy storage device to obtain a target control strategy, wherein the performance adjustment factor is used to adjust an operation performance of the energy storage device, and the operation performance of the energy storage device controlled according to the target control strategy is greater than the operation performance of the energy storage device controlled according to the initial control strategy, and the target control strategy is used to represent a rule for controlling the vehicle to switch the abnormal state; and the execution layer is configured to control the vehicle to switch the operation state of the energy storage device from the abnormal state to a safe state according to the target control strategy at the future time. The system further comprises:

10. The system of claim 9, wherein, The feedback layer is configured to obtain the first state information of the energy storage device in the safe state after the energy storage device is switched from the abnormal state to the safe state, determine a temperature error of the energy storage device based on the first state information, and in response to the temperature error being greater than an error threshold, adjust model parameters of a deep learning model corresponding to the energy storage device and / or model parameters of a long short-term memory network model of the energy storage device. The processor is configured to run a program, and the program is configured to execute the method in any one of claims 1 to 7 when the program is run.

11. A processor, comprising: The computer readable storage medium comprises a stored program, and the program is configured to control a device where the computer readable storage medium is located to execute the method in any one of claims 1 to 7 when the program is run.

12. An electronic device, comprising: The computer program product comprises a computer program, and the computer program is configured to implement the method in any one of claims 1 to 7 when executed by a processor.

13. A computer-readable storage medium, characterized in that, The computer program product comprises a computer program, and the computer program is configured to implement the method in any one of claims 1 to 7 when executed by a processor.

14. A computer program product, characterised in that, The computer program product comprises a computer program, and the computer program is configured to implement the method in any one of claims 1 to 7 when executed by a processor.

15. A vehicle characterized by comprising: The computer program product comprises a computer program, and the computer program is configured to implement the method in any one of claims 1 to 7 when executed by a processor. ​ ​

Images