Battery pack protection method, electronic equipment and vehicle

Abstract

The invention relates to the technical field of vehicle safety, and provides a battery pack protection method, electronic equipment and a vehicle, the current environment information of the vehicle at the current moment and the current vehicle operation information are acquired, and a target wading time sequence vector is determined according to the current environment information and the current vehicle operation information; inputting the target wading time sequence vector into a pre-trained risk determination model, processing the target wading time sequence vector by the risk determination model, and outputting a target risk value; and determining a target risk threshold, determining a target protection strategy according to the target risk value in response to the fact that the target risk value is greater than the target risk threshold, and protecting the vehicle battery pack according to the target protection strategy. The battery pack is protected through the target protection strategy, so that the driving safety is ensured. Meanwhile, the target protection strategy determined based on the target risk value is better matched with the current actual condition, and the practicability and robustness of battery pack wading protection are improved.

Description

Technical Field

[0001] This disclosure relates to the field of vehicle safety technology, and in particular to a battery pack protection method, electronic equipment, and vehicle. Background Technology

[0002] With the rapid development of vehicle technology, vehicles have become an important means of transportation in people's daily lives. When driving, due to weather and road conditions, vehicles often need to wade through water. The water pressure and impacts from debris during wading can damage the sealing components of the battery pack. If water enters the battery pack, it can lead to internal short circuits, high-voltage system leakage, or component corrosion, causing battery failure, vehicle breakdown, or even safety hazards.

[0003] In related technologies, water immersion detection methods often rely on simple threshold judgments, such as triggering protective actions when the water depth exceeds a set value. These methods do not consider dynamic factors such as water flow velocity, pressure difference fluctuations, and environmental trends, which can easily lead to misjudgments or delayed responses, resulting in low accuracy. Summary of the Invention

[0004] In view of this, the purpose of this disclosure is to propose a battery pack protection method, electronic device and vehicle to solve the problem that the current reliance on simple threshold judgment for water wading detection results in misjudgment or delayed response and low judgment accuracy.

[0005] To achieve the above objectives, a first aspect of this disclosure provides a battery pack protection method, the method comprising: Obtain the current environmental information and vehicle operation information of the vehicle at the current moment, and determine the target wading time sequence vector based on the current environmental information and the current vehicle operation information; The target water wading time series vector is input into a pre-trained risk determination model, and the target risk value is output after processing by the risk determination model. A target risk threshold is determined. In response to the target risk value being greater than the target risk threshold, a target protection strategy is determined based on the target risk value, and the vehicle battery pack is protected according to the target protection strategy.

[0006] Specifically, determining the target wading time-series vector based on the current environmental information and the current vehicle operation information includes: The current environmental information and the current vehicle operation information are input into a pre-trained wading state determination model, which processes the wading state determination model and outputs the target wading state. Obtain historical environmental information and historical vehicle operation information within a preset time period before the current moment; input the historical environmental information, historical vehicle operation information, current environmental information, and current vehicle operation information into a pre-trained state change determination model; process the state change determination model and output the target wading state change trend. Based on the target water wading state and the trend of the target water wading state change, the target water wading time sequence vector is determined.

[0007] Specifically, determining the target water wading time-series vector based on the target water wading state and the trend of its change includes: The target wading state and the target wading state change trend are mapped and aligned to obtain the initial wading state vector and the initial wading state change trend vector. The initial wading state vector and the initial wading state change trend vector are normalized to obtain the wading state vector and the wading state change trend vector. The water wading state vector and the water wading state change trend are concatenated to obtain the target water wading time sequence vector.

[0008] Specifically, determining the target risk threshold includes: Obtain the pre-trained threshold determination model; The current environmental information and the target risk value are input into the threshold determination model, and the target risk threshold is obtained through processing by the threshold determination model.

[0009] Specifically, determining the target protection strategy based on the target risk value includes: Based on the target risk value, determine the target risk level corresponding to the target risk value; Based on the current vehicle operation information and the target risk level, a target protection strategy is determined.

[0010] Specifically, the current vehicle operation information includes vehicle speed, and determining the target protection strategy based on the current vehicle operation information and the target risk level includes: In response to the vehicle speed being greater than a first preset speed threshold and the target risk level being low, the target protection strategy is determined to be to close the vehicle battery pack vents.

[0011] Specifically, the current vehicle operation information includes vehicle speed, and determining the target protection strategy based on the current vehicle operation information and the target risk level includes: In response to the vehicle speed being less than a second preset speed threshold and the target risk level being low, the target protection strategy is determined to be to activate the vehicle battery pack's heating and pressure-maintaining mode.

[0012] Specifically, after determining the target risk threshold, the process also includes: In response to the target risk value being less than or equal to the target risk threshold, it is determined whether there is a risk of water-related incidents based on the current environmental information; In response to the risk of water immersion, a conservative protection mode is activated to protect the vehicle battery pack.

[0013] Based on the same inventive concept, a second aspect of this disclosure proposes a battery pack protection device, comprising: The data acquisition module is configured to acquire the current environmental information and current vehicle operation information of the vehicle at the current moment, and determine the target wading time sequence vector based on the current environmental information and current vehicle operation information; The water wading risk determination module is configured to input the target water wading time series vector into a pre-trained risk determination model, process it through the risk determination model, and output the target risk value. A battery protection module is configured to determine a target risk threshold, and in response to the target risk value being greater than the target risk threshold, determine a target protection strategy based on the target risk value, and protect the vehicle battery pack according to the target protection strategy.

[0014] Based on the same inventive concept, a third aspect of this disclosure proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the battery pack protection method as described above when executing the computer program.

[0015] Based on the same inventive concept, a fourth aspect of this disclosure provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the battery pack protection method as described above.

[0016] Based on the same inventive concept, the fifth aspect of this disclosure provides a vehicle including the battery pack protection device described in the second aspect, the electronic device described in the third aspect, or the storage medium described in the fourth aspect.

[0017] As can be seen from the above, this disclosure proposes a battery pack protection method, electronic device, and vehicle. It acquires the current environmental information and vehicle operation information of the vehicle at the current moment, and determines a target wading time-series vector based on the current environmental information and the current vehicle operation information. The target wading time-series vector describes the changes in the wading environment, thus considering the dynamic nature of state changes during wading when determining wading risk based on the target wading time-series vector, resulting in a more accurate assessment of battery pack risk. The target wading time-series vector is input into a pre-trained risk determination model, which processes the data and outputs a target risk value. Because the risk determination model is trained on a large amount of data, the target risk value of the battery pack determined using the trained risk determination model is more accurate. A target risk threshold is determined. In response to the target risk value being greater than the target risk threshold, a target protection strategy is determined based on the target risk value, and the vehicle battery pack is protected according to the target protection strategy. When the target risk value exceeds the target risk threshold, it indicates a significant threat to the battery pack's safety during vehicle wading. Therefore, a corresponding target protection strategy is determined. This strategy protects the battery pack and ensures driving safety. Furthermore, the target protection strategy, based on the target risk value, is more closely aligned with the actual situation, improving the practicality and robustness of the battery pack's water wading protection. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this disclosure or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of a battery pack protection method according to an embodiment of the present disclosure; Figure 2 This is a structural block diagram of the battery pack protection device according to an embodiment of the present disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0021] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0022] The following are definitions of terms used in this disclosure: SimCLR: SimCLR (Simple Framework for Contrastive Learning of Visual Representations) is a self-supervised visual representation learning framework. Its core idea is to enable the model to learn to distinguish between similar and dissimilar samples in unlabeled images through contrastive learning. It utilizes data augmentation (such as cropping and color perturbation) to generate different views of the same image, extracts features through an encoder, and maps them to a contrast space using a non-linear projection head. This maximizes the consistency between different views of the same sample while minimizing its similarity to other samples.

[0023] BiLSTM: Bi-LSTM (Bidirectional Long Short-Term Memory Network) concatenates two LSTM layers one after the other: one reads the sequence in forward chronological order and the other reads the sequence in reverse chronological order, extracting contextual features from the past and future respectively, and then concatenates or adds the two sets of hidden states bit by bit to obtain a representation that simultaneously perceives the global context.

[0024] One-Class SVM: One-Class SVM is an unsupervised anomaly detection algorithm that is trained using only "normal" samples to find an optimal hyperplane that encloses most of the normal data. New samples that fall outside the hyperplane are judged as anomalies. Compared to multi-class SVM, it does not require labeling anomalies and can work with only one class of data.

[0025] CAN signal: The vehicle CAN signal is a serial data frame transmitted through two differential lines (CAN_H and CAN_L) in a "broadcast + priority arbitration" manner. Each frame is identified by a unique identifier indicating its content and priority. All ECUs listen to the signal simultaneously and only receive messages that are relevant to themselves, so as to realize real-time and reliable communication between systems such as engine, transmission, and brake.

[0026] PWM signal: The vehicle PWM signal is a square wave that uses a fast switching circuit to "adjust the duty cycle and stabilize the average voltage". The ECU uses it to precisely control actuators such as motors, solenoid valves, and lights, just like a dimmer knob. Only digital high / low signals are transmitted in the wiring harness, which has strong anti-interference and high efficiency.

[0027] With the rapid development of vehicle technology, vehicles have become an important means of transportation in people's daily lives. When driving, due to weather and road conditions, vehicles often need to wade through water. The water pressure and impacts from debris during wading can damage the sealing components of the battery pack. If water enters the battery pack, it can lead to internal short circuits, high-voltage system leakage, or component corrosion, causing battery failure, vehicle breakdown, or even safety hazards.

[0028] In related technologies, water immersion detection methods often rely on simple threshold judgments, such as triggering protective actions when the water depth exceeds a set value. These methods do not consider dynamic factors such as water flow velocity, pressure difference fluctuations, and environmental trends, which can easily lead to misjudgments or delayed responses, resulting in low accuracy.

[0029] Meanwhile, current methods generally rely on manually labeled data to build classification models, making it difficult to adapt to real-world wading environments characterized by unlabeled, large-scale, and variable road conditions. Due to the high noise levels in sensor signals and the rapid evolution of abnormal states in wading conditions, the lack of in-depth modeling of temporal changes makes it difficult for traditional detection systems to identify potential risks in a timely and accurate manner.

[0030] The root cause of these problems lies in the fact that the relevant technologies have neglected the dynamic, multidimensional and uncertain nature of state changes during water wading, and lack an adaptive learning recognition mechanism and a systematically integrated intelligent protection strategy, which seriously restricts the practicality and robustness of battery pack water wading protection systems in complex real vehicle environments.

[0031] Therefore, to solve the above problems, this embodiment proposes a battery pack protection method, such as... Figure 1 As shown, the method includes: Step 101: Obtain the current environmental information and current vehicle operation information of the vehicle at the current moment, and determine the target wading time sequence vector based on the current environmental information and the current vehicle operation information.

[0032] In specific implementation, the current environmental information and vehicle operation information of the vehicle at the current moment are obtained. The current environmental information includes water level information and water flow velocity. The current vehicle operation information includes at least one of the following: external air pressure of the battery pack, internal air pressure of the battery pack, vehicle acceleration, vehicle angular velocity, vehicle position, and vehicle speed. The water flow velocity includes lateral water flow velocity and longitudinal water flow velocity.

[0033] In this embodiment, the current environmental information and current vehicle operation information are data collected by various sensors inside the vehicle, such as real-time water level information collected by a water depth sensor, lateral or longitudinal water flow velocity obtained by a water flow velocity sensor, air pressure change signals between the outside and inside of the battery pack obtained by a differential pressure sensor, acceleration and angular velocity information recorded by an IMU module, geographical location information provided by a GPS module, and basic vehicle status data such as vehicle speed sensor and braking status sensor. These multi-source signals together constitute the basic data input required for water wading risk identification.

[0034] After obtaining the current environmental information and the current vehicle operation information, a target wading time-series vector is determined based on the current environmental information and the current vehicle operation information. This target wading time-series vector describes the vehicle's current wading state and the trend of future wading state changes; that is, the target wading time-series vector describes the wading environment change information.

[0035] Step 102: Input the target water wading time series vector into the pre-trained risk determination model, process it through the risk determination model, and output the target risk value.

[0036] In practice, a pre-trained risk determination model is obtained, the target water wading time series vector is input into the risk determination model, and after processing by the risk determination model, the target risk value is finally output.

[0037] In this embodiment, the training process of the risk determination model specifically includes: Step a, obtain the first training dataset and the initial risk determination model, wherein the first training dataset includes historical wading time series vectors and historical risk values.

[0038] Step b: Input the training data from the first training dataset into the initial risk determination model for training, and determine that the first preset training termination condition is met to obtain the risk determination model.

[0039] In specific implementation, a first training dataset and an initial risk determination model are obtained, wherein the first training dataset includes historical water wading time series vectors and historical risk values. Training data from the first training dataset is input into the initial risk determination model for training. Once a first preset training termination condition is met, the risk determination model is obtained.

[0040] The first preset training termination condition includes at least one of the following: determining that all data in the first training dataset has been input into the initial risk determination model for training, determining that the loss function of the initial risk determination model has converged to a first convergence threshold, or determining that the initial risk determination model has been iterated for training to a first preset number of iterations.

[0041] For example, the first preset training termination condition is that all data in the first training dataset has been input into the initial risk determination model for training: The first training dataset contains fifty sets of data, each set including a historical water wading time-series vector and a historical risk value. The first preset training termination condition is that all data in the first training dataset has been input into the initial risk determination model for training. That is, when all fifty sets of data have been input into the initial risk determination model, there is no training data in the first training dataset that has not yet been input into the initial risk determination model. At this point, the initial risk determination model training is considered complete, and the risk determination model is obtained.

[0042] In another example, the first preset training termination condition is that the loss function of the model determines the initial risk and converges to a first convergence threshold: The training data from the first training dataset is input into the initial risk determination model for training, and the training results are output. A loss function is determined based on the training results and the historical fault types, historical fault handling methods, and historical fault causes. The loss function may include at least one of the following: mean squared error loss function, cross-entropy loss function, logarithmic loss function, exponential loss function, squared loss function, or absolute value loss function, etc. When the loss function converges to a first convergence threshold, it is determined that a first preset training termination condition has been met, and the risk determination model is obtained.

[0043] In another example, the first preset training termination condition is to determine the initial risk and then iterate the model training to the first preset number of iterations.

[0044] The training data in the first training dataset is input into the initial risk determination model for iterative training. The number of iterations is recorded. When the number of iterations is equal to the first preset number of iterations, the first preset training termination condition is met, and the risk determination model is obtained.

[0045] The above scheme uses training data in the first training dataset, including historical wading time-series vectors and historical risk values. The initial vehicle fault analysis model is trained using the training data in the first training dataset until the first preset training termination condition is met, resulting in a risk determination model. This model is then used to output the target risk value corresponding to the target wading time-series vector, making the determined target risk value more accurate.

[0046] Step 103: Determine the target risk threshold; in response to the target risk value being greater than the target risk threshold, determine the target protection strategy based on the target risk value, and protect the vehicle battery pack according to the target protection strategy.

[0047] In practice, a target risk threshold is determined, and the target risk value is compared with the target risk threshold. If the target risk value is greater than the target risk threshold, it indicates abnormal water wading behavior. To avoid damage to the vehicle battery pack, protective measures must be taken to protect the battery pack.

[0048] Different target risk values ​​correspond to different levels of water wading risk, and consequently, different levels of threat to the battery pack, thus requiring different protection strategies. Therefore, a target protection strategy is determined based on the target risk value, and the vehicle battery pack is protected according to the target protection strategy.

[0049] The above scheme obtains the current environmental information and vehicle operation information of the vehicle at the current moment, and determines the target wading time-series vector based on the current environmental information and vehicle operation information. The target wading time-series vector describes the changes in the wading environment, thus considering the dynamic nature of state changes during wading when determining wading risk, resulting in a more accurate assessment of battery pack risk. The target wading time-series vector is input into a pre-trained risk determination model, which processes the data and outputs a target risk value. Because the risk determination model has been trained on a large amount of data, the target risk value of the battery pack determined using the trained model is more accurate. A target risk threshold is determined. If the target risk value is greater than the target risk threshold, a target protection strategy is determined based on the target risk value, and the vehicle battery pack is protected according to the target protection strategy. When the target risk value is greater than the target risk threshold, it indicates a significant threat to the battery pack's safety during the vehicle's current wading journey. Therefore, a corresponding target protection strategy is determined at this time to protect the battery pack and ensure driving safety. On the other hand, the target protection strategy determined based on the target risk value is more in line with the current situation, improving the practicality and robustness of water immersion protection for battery packs.

[0050] In some embodiments, after collecting raw data corresponding to current environmental information and current vehicle operation information through various sensors inside the vehicle, the raw data is preprocessed to obtain current environmental information and current vehicle operation information, and multi-dimensional time series data is constructed based on the current environmental information and current vehicle operation information. Subsequently, when determining the target wading time series vector, the multi-dimensional time series data is used for determination.

[0051] In this embodiment, the original data is processed by a data processing model to obtain multidimensional time series data. The data processing model includes a synchronization and time alignment module, a signal noise reduction module, and a feature standardization and completion module. The specific process includes: The raw data is input into the synchronization and time alignment module, and a unified timestamp mechanism is used to align the various types of raw data in time, ensuring that the multi-dimensional signals of water depth, water flow velocity, pressure difference and vehicle status remain consistent in each time window.

[0052] The aligned data is input to the signal noise reduction module to filter out data segments with drastic jumps or obvious errors. The main approach is to use a sliding window statistical strategy and upper and lower limit methods to remove erroneous readings caused by factors such as mud splashing, sensor jitter, and electrical interference.

[0053] Finally, the feature standardization and completion module is used to perform linear interpolation to complete segments of the signal that have short-term acquisition loss or abnormal interruption, and the features of all dimensions are normalized and standardized to ensure that the data output from different sensors have uniform scale and distribution characteristics. This results in structured, unified-format multidimensional time-series data, where each time window corresponds to a complete set of water-related sensing features.

[0054] The above scheme processes the raw data to ensure the stability and robustness of the input data when determining the target water-related time-series vector, thereby meeting the real-time and accuracy requirements for water-related risk identification.

[0055] In some embodiments, the target wading time-series vector describes the vehicle's current wading state and the trend of future wading state changes. Therefore, the wading state and the trend of future wading state changes are determined based on the vehicle's current environmental information and vehicle operation information, respectively. That is, determining the target wading time-series vector based on the current environmental information and the current vehicle operation information in step 101 specifically includes: Step 1011: Input the current environmental information and the current vehicle operation information into the pre-trained wading state determination model, process the wading state determination model, and output the target wading state. Step 1012: Obtain historical environmental information and historical vehicle operation information within a preset time period before the current moment; input the historical environmental information, historical vehicle operation information, current environmental information, and current vehicle operation information into a pre-trained state change determination model; process the state change determination model and output the target wading state change trend. Step 1013: Determine the target wading time sequence vector based on the target wading state and the trend of the target wading state change.

[0056] In practice, a pre-trained wading state determination model is acquired, and then the current environmental information and the current vehicle operation information are input into the wading state determination model. The model processes the data and outputs the target wading state. The target wading state represents the wading behavior state.

[0057] In this embodiment, the water wading status determination model adopts a neural network structure. To reduce data labeling costs, the water wading status determination model in this embodiment is a self-supervised contrastive learning model, namely the SimCLR model. A stable and highly discriminative temporal feature encoder is trained using the SimCLR self-supervised contrastive learning mechanism to extract feature representations capable of discriminating water wading risks from a large amount of unlabeled water wading data without manual labeling.

[0058] Specifically, the internal structure of the wading state determination model includes a data augmentation module, an encoder network, a contrast view generation module, and a feature representation optimization module. Taking the construction of multi-dimensional time series data based on current environmental information and current vehicle operation information as an example, the process of determining the target wading state specifically includes: The data augmentation module generates two different perturbation versions for each set of multidimensional time series data to simulate the different sensor performances of vehicles under similar wading conditions. The augmentation strategies include randomly blocking signals for part of the time period, adding noise interference, randomly stretching the water depth curve, and simulating sudden changes in water flow velocity. Through different augmentation strategies, two sets of deformed samples that still represent the same potential wading conditions are formed.

[0059] Two sets of enhanced samples are fed into the encoder network for feature extraction. The encoder structure consists of temporal convolution and gated recurrent units, which can extract local change patterns while retaining long-term trend information, and finally output a fixed-length feature vector.

[0060] Based on the SimCLR framework, the comparison view generation module generates two enhanced versions of the same original sample as positive sample pairs and negative sample pairs between different samples. In the feature space, the encoder is optimized by bringing the positive sample pairs closer and the negative sample pairs further apart, so that the model learns to capture stable structural features during the wading process.

[0061] The feature representation optimization module continuously adjusts the encoder parameters using this unsupervised comparison method, enabling the output features to more accurately reflect the similarities and differences between wading states. This ultimately forms a feature extractor capable of encoding arbitrary temporal sensor data into low-dimensional vectors. This feature extractor can extract consistent discriminative representations across different wading scenarios, effectively improving the overall model's generalization ability and environmental adaptability.

[0062] The final output is a set of fixed-length temporal encoded vectors, each corresponding to a description of the wading state within a time window, thus obtaining the target wading state. This vector serves as the direct input for subsequent anomaly identification and trend modeling. The entire SimCLR training process does not rely on any labels; it completes representation learning solely through a mechanism that compares the distance between the original data augmentation and the feature space, greatly reducing data annotation costs while exhibiting strong robustness and transferability.

[0063] The system acquires historical environmental information and historical vehicle operation information within a preset time period prior to the current moment. It then inputs the historical environmental information, historical vehicle operation information, current environmental information, and current vehicle operation information into a pre-trained state change determination model. Based on the state change determination model, the system processes the data and outputs the target wading state change trend.

[0064] In this embodiment, the state change determination model employs a neural network structure. To enhance the ability to identify scenarios with slow changes in water wading conditions, this embodiment uses a BiLSTM model that integrates time-series features. By introducing time-series modeling capabilities, the evolution pattern of water wading conditions can be understood over a longer time domain, supplementing the static feature expression model's lack of perception of historical trend information and improving the ability to judge potential risk escalation processes.

[0065] Specifically, the internal structure of the state change determination model includes a historical data construction module, a bidirectional recurrent neural network modeling module, and a state trend vector extraction module. Taking the construction of multi-dimensional time series data based on current environmental information and current vehicle operation information as an example, the process of determining the target wading state change trend specifically includes: The historical data construction module extracts multi-dimensional sensor data fragments from historical environmental information and historical vehicle operation information over a fixed time window strategy to obtain historical time-series data. The commonly used window length is set to 5 to 10 seconds, covering the entire process of the initial, middle, and impending risk stages of wading, ensuring that the model can capture the slow evolution of water flow intensification, water depth rise, or sudden pressure changes.

[0066] The historical time series data and multidimensional time series data are input into the state trend vector extraction module for modeling. At the same time, the module extracts context state evolution features from the forward sequence from the past to the present and the reverse sequence from the present to the past, thus possessing the ability to perceive time dependencies in a comprehensive manner.

[0067] During the training phase of the state change determination model, the prediction error of the signal value at the next time step is used as the training objective, eliminating the need for manual labels and thus constructing a stable self-supervised time trend modeler. After training, the output of the state change determination model is a trend vector containing complete contextual dynamic information, namely the target wading state change trend, which represents the evolution trajectory and development direction of the vehicle's current wading state.

[0068] Because BiLSTM networks inherently possess strong sequence modeling capabilities, they are suitable for identifying scenarios where wading conditions change slowly but are extremely dangerous. Their training method in a label-free environment exhibits good scalability and practical deployability. Ultimately, the trend of wading condition changes will serve as a crucial supplementary signal in the anomaly identification process, not only improving the overall sensitivity of the identification model but also enhancing its ability to cope with slowly accumulating risks, thus ensuring the system's forward-looking judgment level in wading situations.

[0069] Based on the target wading state and the trend of its change, a target wading time-series vector is determined. The method for determining the target wading time-series vector specifically includes: Step A: Map and align the target wading state and the target wading state change trend to obtain the initial wading state vector and the initial wading state change trend vector. Step B: Normalize the initial wading state vector and the initial wading state change trend vector to obtain the wading state vector and the wading state change trend vector. Step C involves concatenating the wading state vector and the wading state change trend vector to obtain the target wading time sequence vector.

[0070] In specific implementation, a linear transformation is performed on the target wading state to map it to a preset semantic space, resulting in an initial wading state vector. A linear transformation is then performed on the target wading state change trend to map it to the preset semantic space, resulting in an initial wading state change trend vector.

[0071] The initial wading state vector and the initial wading state change trend vector are normalized to obtain the wading state vector and the wading state change trend vector. The wading state vector and the wading state change trend vector are then concatenated to obtain the target wading time series vector.

[0072] The above scheme obtains the target wading time-series vector by concatenating the wading state vector and the wading state change trend. The resulting target wading time-series vector simultaneously possesses the current state expression provided by the target wading state and the trend prediction capability provided by the wading state change trend vector, which greatly improves the timeliness and accuracy of anomaly detection.

[0073] In some embodiments, to improve adaptability to uncertainties such as different terrains, climates, and driving styles, the risk threshold can be dynamically adjusted when determining whether abnormal wading situations exist. Specifically, determining the target risk threshold as described in step 103 includes: Step 1031: Obtain the pre-trained threshold determination model; Step 1032: Input the current environmental information and the target risk value into the threshold determination model, and obtain the target risk threshold through the threshold determination model.

[0074] In practice, a pre-trained threshold determination model is obtained, and the current environmental information and the target risk value are input into the threshold determination model. The target risk threshold is obtained through processing by the threshold determination model. The environmental information includes dynamic environmental indicators such as current vehicle speed, water depth, wading duration, road slope estimation, and pressure difference change rate.

[0075] In this embodiment, the threshold determination model is a Bayesian estimation model, which includes a state perception module, a prior probability generation module, a posterior risk score estimation module, and a risk threshold adjuster.

[0076] The state perception module is responsible for aggregating vehicle dynamic parameters and environmental perception information in real time, using them as contextual factors in the subsequent probabilistic modeling process. Typical inputs include complex combinations such as rapidly increasing water depth without pressure differential response and high vehicle speed with abnormal water flow velocity. The prior probability generation module generates initial risk distributions for different wading scenarios based on historical operating condition statistics or risk models provided by the manufacturer, providing a foundation for Bayesian updates. The posterior risk score estimation module calculates the posterior probability of actual wading risk under the current operating condition by combining the anomaly score given by the current fusion model with the prior probability corresponding to the vehicle's current state, reflecting the system's awareness of uncertainty in risk judgment. The risk threshold adjuster automatically adjusts the risk judgment threshold according to the changing trend of the posterior probability, ensuring the model maintains sufficient tolerance in low-speed, shallow water environments to avoid false triggers, while significantly improving risk sensitivity in dangerous scenarios such as high-speed, deep water or strong currents, achieving priority enhancement and improved response timeliness. This not only improves the adaptability of risk judgment but also enhances the model's adaptability to uncertain factors such as different terrains, climates, and driving styles, giving the wading risk identification system stronger environmental generalization ability and practicality.

[0077] In this embodiment, the risk prior model can be continuously optimized as real vehicle data accumulates, improving the accuracy and precision of dynamic threshold adjustment and meeting the safety protection needs of new energy vehicles under complex off-road and wading conditions.

[0078] The above scheme constructs a risk threshold adjustment mechanism using a Bayesian estimation framework. This mechanism adjusts the sensitivity of anomaly detection in real time under different vehicle conditions and external environmental conditions, avoiding misjudgment problems caused by fixed thresholds and achieving more accurate water wading risk identification. At the same time, it ensures that the vehicle system avoids overreaction or frequent false triggers while ensuring safety.

[0079] In some embodiments, when determining the target protection strategy, the current water risk level and vehicle operation information can be comprehensively considered. Specifically, determining the target protection strategy based on the target risk value in step 103 includes: Step 103A: Determine the target risk level corresponding to the target risk value based on the target risk value; Step 103B: Determine the target protection strategy based on the current vehicle operation information and the target risk level.

[0080] In practice, a target risk level is determined based on the target risk value. Specifically, the target risk value can be searched in a database to determine the target risk level. The database stores the correspondence between risk values ​​and risk levels, and the target risk level represents the risk level of the vehicle currently wading through water, i.e., the degree of threat posed by water to the vehicle's battery pack. In this embodiment, the target risk level specifically includes low, medium, and high levels.

[0081] Based on the current vehicle operating information and the target risk level, a target protection strategy is determined. This target protection strategy is a battery pack protection control command set, and specific actions include at least one of the following: automatically closing the upper and lower ventilation openings of the battery pack, switching to battery sealed protection mode, triggering the battery heating device to maintain internal positive pressure, disconnecting the external charging interface, etc., and broadcasting these commands to the relevant execution modules via the vehicle controller network to complete the protection actions.

[0082] Specifically, the current vehicle operation information includes vehicle speed, and determining the target protection strategy based on the current vehicle operation information and the target risk level includes: Step 10A: In response to the vehicle speed being greater than a first preset speed threshold and the target risk level being low, the target protection strategy is determined to be to close the vehicle battery pack vents.

[0083] In practice, the vehicle speed is compared with a first preset speed threshold. If the vehicle speed is greater than the first preset speed threshold and the target risk level is low (i.e., the vehicle is traveling at high speed and the risk of wading is low), the target protection strategy is determined to be closing the vehicle battery pack ventilation openings. The first preset speed threshold is a relatively high vehicle speed, such as 70 km / h.

[0084] Based on the above approach, if the vehicle is traveling at high speed and the risk of wading through water is low, the target protection strategy is to close the vehicle battery pack ventilation openings instead of immediately disconnecting the vehicle's high-voltage system. This avoids directly disconnecting the high-voltage system from threatening the normal operation of the vehicle and reducing driving safety.

[0085] Specifically, the current vehicle operation information includes vehicle speed, and determining the target protection strategy based on the current vehicle operation information and the target risk level includes: Step 10B: In response to the vehicle speed being less than the second preset speed threshold and the target risk level being low, the target protection strategy is determined to be to activate the heating and pressure-maintaining mode of the vehicle battery pack.

[0086] In specific implementation, the vehicle speed is compared with a second preset speed threshold. If the vehicle speed is less than the second preset speed threshold and the target risk level is low (i.e., the vehicle is traveling at an extremely low speed), the target protection strategy is determined to be to activate the vehicle battery pack's heating and pressure-maintaining mode to prevent moisture backflow. The second preset speed threshold is an extremely low vehicle speed value, such as 5 km / h or 4 km / h. Preferably, in this embodiment, the second preset speed threshold is 0.

[0087] With the above solution, if the vehicle is traveling at extremely low speeds, such as when wading in still water, it will first enter the heating and pressure-maintaining mode to prevent water vapor backflow.

[0088] In some embodiments, after determining the target protection strategy, the strategy is concretized into executable CAN or PWM signals, which are then used by subsystems such as the battery pack thermal management controller, vent electric actuators, and charging management module to execute corresponding actions. Simultaneously, the response status of each execution unit is collected in real time and fed back to the central controller to form a closed loop, ensuring that the execution of each protection action can be confirmed and that redundant remedial channels are available in abnormal situations. Judgment and issuance are completed within milliseconds, ensuring the system has rapid response capabilities in the event of sudden flooding or a rapid rise in water level.

[0089] In this embodiment, the protection level and triggering conditions can be flexibly configured through vehicle control parameters during the vehicle development phase, and it has remote OTA online update capabilities, facilitating the synchronous optimization of control strategies as the risk identification model is upgraded. Simultaneously, each protective action and its corresponding risk identification data are uploaded to the cloud for subsequent modeling training and strategy optimization, thus forming a data-driven, continuous safety enhancement closed-loop system. By constructing a stable bridge between model recognition and the vehicle execution system, intelligent battery pack protection for new energy off-road vehicles under wading conditions is achieved, effectively improving the vehicle's survivability and system robustness in extreme environments.

[0090] In some embodiments, after determining the target risk threshold, if the target risk value does not exceed the target risk threshold, it can be further determined whether there is a risk of water intrusion based on the current environmental information. That is, after determining the target risk threshold in step 103, the method further includes: Step 10a: In response to the target risk value being less than or equal to the target risk threshold, determine whether there is a risk of water intrusion based on the current environmental information; Step 10b: In response to the risk of water immersion, a conservative protection mode is activated to protect the vehicle battery pack based on the conservative protection mode.

[0091] In practice, after determining the target risk threshold, the target risk value is compared with the target risk threshold. If the target risk value is less than or equal to the target risk threshold, then the existence of water-related risk is determined based on the current environmental information, specifically including: The acquired current environmental information is compared with an environmental data threshold. For example, the current environmental information includes water depth information, which is then compared with a water depth threshold. If the water depth information is greater than the water depth threshold, an abnormal water depth is detected by the sensor.

[0092] Based on the current environmental information indicating a risk of water damage, a conservative protection mode is activated to protect the vehicle's battery pack. This conservative protection mode represents the most basic strategy for protecting the vehicle's battery pack, such as disconnecting the external charging port and displaying warning messages on the dashboard.

[0093] The above scheme configures a hard timeout logic for risk assessment. If the model does not make a clear judgment within the set time but the sensor detects an abnormal state, it will automatically switch to a conservative protection mode to avoid control failure due to model delay or sensor failure.

[0094] It should be noted that the method of this disclosure embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this disclosure embodiment, and the multiple devices will interact with each other to complete the method described.

[0095] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] Based on the same inventive concept, corresponding to any of the above embodiments, this disclosure also provides a battery pack protection device.

[0097] refer to Figure 2 , Figure 2 The battery pack protection device of the embodiment includes: The data acquisition module 201 is configured to acquire the current environmental information and current vehicle operation information of the vehicle at the current moment, and determine the target wading time sequence vector based on the current environmental information and current vehicle operation information; The water wading risk determination module 202 is configured to input the target water wading time series vector into a pre-trained risk determination model, process it through the risk determination model, and output the target risk value. The battery protection module 203 is configured to determine a target risk threshold, and in response to the target risk value being greater than the target risk threshold, determine a target protection strategy based on the target risk value, and protect the vehicle battery pack according to the target protection strategy.

[0098] In some embodiments, the data acquisition module 201 is specifically configured as follows: The current environmental information and the current vehicle operation information are input into a pre-trained wading state determination model, which processes the wading state determination model and outputs the target wading state. Obtain historical environmental information and historical vehicle operation information within a preset time period before the current moment; input the historical environmental information, historical vehicle operation information, current environmental information, and current vehicle operation information into a pre-trained state change determination model; process the state change determination model and output the target wading state change trend. Based on the target water wading state and the trend of the target water wading state change, the target water wading time sequence vector is determined.

[0099] In some embodiments, the data acquisition module 201 is specifically configured as follows: The target wading state and the target wading state change trend are mapped and aligned to obtain the initial wading state vector and the initial wading state change trend vector. The initial wading state vector and the initial wading state change trend vector are normalized to obtain the wading state vector and the wading state change trend vector. The water wading state vector and the water wading state change trend are concatenated to obtain the target water wading time sequence vector.

[0100] In some embodiments, the battery protection module 203 is specifically configured as follows: Obtain the pre-trained threshold determination model; The current environmental information and the target risk value are input into the threshold determination model, and the target risk threshold is obtained through processing by the threshold determination model.

[0101] In some embodiments, the battery protection module 203 is specifically configured as follows: Based on the target risk value, determine the target risk level corresponding to the target risk value; Based on the current vehicle operation information and the target risk level, a target protection strategy is determined.

[0102] In some embodiments, the current vehicle operating information includes vehicle speed, and the battery protection module 203 is specifically configured as follows: In response to the vehicle speed being greater than a first preset speed threshold and the target risk level being low, the target protection strategy is determined to be to close the vehicle battery pack vents.

[0103] In some embodiments, the current vehicle operating information includes vehicle speed, and the battery protection module 203 is specifically configured as follows: In response to the vehicle speed being less than a second preset speed threshold and the target risk level being low, the target protection strategy is determined to be to activate the vehicle battery pack's heating and pressure-maintaining mode.

[0104] In some embodiments, the battery protection module 203 is specifically configured as follows: In response to the target risk value being less than or equal to the target risk threshold, it is determined whether there is a risk of water-related incidents based on the current environmental information; In response to the risk of water immersion, a conservative protection mode is activated to protect the vehicle battery pack.

[0105] For ease of description, the above apparatus is described in terms of its functions, divided into various modules. Of course, in implementing this disclosure, the functions of each module can be implemented in one or more software and / or hardware.

[0106] The apparatus of the above embodiments is used to implement the corresponding battery pack protection method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0107] Based on the same inventive concept, corresponding to any of the above embodiments, this disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the battery pack protection method described in any of the above embodiments.

[0108] Figure 3 This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0109] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0110] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0111] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0112] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0113] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0114] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0115] The electronic devices described above are used to implement the corresponding battery pack protection methods in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0116] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this disclosure also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the battery pack protection method as described in any of the above embodiments.

[0117] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0118] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the battery pack protection method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0119] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a vehicle, including the battery pack protection device in the above embodiments, the electronic device in the above embodiments, and the computer-readable storage medium in the above embodiments, wherein the vehicle device implements the battery pack protection method described in any of the above embodiments.

[0120] The vehicles described in the above embodiments are used to implement the battery pack protection method described in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0121] It is understood that before using the technical solutions of the various embodiments in this disclosure, users will be informed of the type, scope of use, and usage scenarios of the personal information involved in an appropriate manner, and user authorization will be obtained.

[0122] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose, based on the prompt message, whether to provide personal information to the software or hardware such as electronic devices, applications, servers, or storage media performing the operations of this disclosed technical solution.

[0123] As an optional but not limited implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0124] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0125] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this disclosure (including the claims) is limited to these examples; within the framework of this disclosure, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this disclosure as described above, which are not provided in detail for the sake of brevity.

[0126] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this disclosure, the provided drawings may or may not show well-known power / ground connections to integrated circuit (IC) chips and other components. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this disclosure, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this disclosure will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this disclosure, it will be apparent to those skilled in the art that the embodiments of this disclosure can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0127] Although this disclosure has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0128] This disclosure is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A battery pack protection method, characterized in that, include: Obtain the current environmental information and vehicle operation information of the vehicle at the current moment, and determine the target wading time sequence vector based on the current environmental information and the current vehicle operation information; The target water wading time series vector is input into a pre-trained risk determination model, and the target risk value is output after processing by the risk determination model. A target risk threshold is determined. In response to the target risk value being greater than the target risk threshold, a target protection strategy is determined based on the target risk value, and the vehicle battery pack is protected according to the target protection strategy.

2. The method according to claim 1, characterized in that, The step of determining the target wading time-series vector based on the current environmental information and the current vehicle operation information includes: The current environmental information and the current vehicle operation information are input into a pre-trained wading state determination model, which processes the wading state determination model and outputs the target wading state. Obtain historical environmental information and historical vehicle operation information within a preset time period before the current moment; input the historical environmental information, historical vehicle operation information, current environmental information, and current vehicle operation information into a pre-trained state change determination model; process the state change determination model and output the target wading state change trend. Based on the target water wading state and the trend of the target water wading state change, the target water wading time sequence vector is determined.

3. The method according to claim 2, characterized in that, The step of determining the target water wading time-series vector based on the target water wading state and the trend of its change includes: The target wading state and the target wading state change trend are mapped and aligned to obtain the initial wading state vector and the initial wading state change trend vector. The initial wading state vector and the initial wading state change trend vector are normalized to obtain the wading state vector and the wading state change trend vector. The water wading state vector and the water wading state change trend are concatenated to obtain the target water wading time sequence vector.

4. The method according to claim 1, characterized in that, The determination of the target risk threshold includes: Obtain the pre-trained threshold determination model; The current environmental information and the target risk value are input into the threshold determination model, and the target risk threshold is obtained through processing by the threshold determination model.

5. The method according to claim 1, characterized in that, The step of determining the target protection strategy based on the target risk value includes: Based on the target risk value, determine the target risk level corresponding to the target risk value; Based on the current vehicle operation information and the target risk level, a target protection strategy is determined.

6. The method according to claim 5, characterized in that, The current vehicle operating information includes vehicle speed. The step of determining the target protection strategy based on the current vehicle operation information and the target risk level includes: In response to the vehicle speed being greater than a first preset speed threshold and the target risk level being low, the target protection strategy is determined to be to close the vehicle battery pack vents.

7. The method according to claim 5, characterized in that, The current vehicle operating information includes vehicle speed. The step of determining the target protection strategy based on the current vehicle operation information and the target risk level includes: In response to the vehicle speed being less than a second preset speed threshold and the target risk level being low, the target protection strategy is determined to be to activate the vehicle battery pack's heating and pressure-maintaining mode.

8. The method according to claim 1, characterized in that, After determining the target risk threshold, the following is also included: In response to the target risk value being less than or equal to the target risk threshold, it is determined whether there is a risk of water intrusion based on the current environmental information; In response to the risk of water immersion, a conservative protection mode is activated to protect the vehicle battery pack.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in any one of claims 1 to 8.

10. A vehicle, characterized in that, The vehicle includes the electronic equipment as described in claim 9.

Images