Online safety management and early warning method for power battery of electric vehicle

Abstract

The invention discloses an online safety management and early warning method for a power battery of an electric vehicle, which comprises the following steps of: acquiring multi-dimensional operation data of the power battery and preprocessing the multi-dimensional operation data to obtain a standardized data set; obtaining a dynamic working condition and a time sequence feature based on the standardized data set, generating a battery state feature set and judging whether the battery state feature set meets a preset condition, and if not, optimizing the battery state feature set to obtain an optimized feature set; constructing a risk assessment model, and combining the optimized feature set to obtain a battery safety risk probability and judge a risk level; if the risk level exceeds a threshold value, obtaining a personalized management strategy generated by the cloud based on the optimized feature set and issuing the personalized management strategy to a vehicle end controller; and fusing vehicle end real-time data and battery aging data through an online learning algorithm, dynamically adjusting risk assessment model parameters, generating an adaptive model, and outputting a graded early warning signal. According to the method, the safety risk of the power battery under the complex working condition is effectively evaluated, and a real-time early warning and personalized management strategy is provided.

Description

Technical Field

[0001] The present invention belongs to the field of information technology, and particularly relates to an online safety management and warning method for power batteries of electric vehicles. Background Technique

[0002] As the core direction of new energy vehicles, the safety of the power batteries of electric vehicles is directly related to the reliability of vehicle operation and the safety of users' lives and property. Power battery management is not only a key fulcrum for the development of electric vehicle technology but also an important guarantee for promoting green transportation and energy transformation. However, the current power battery safety management methods have significant limitations in practical applications. Many solutions rely too much on a single model and are difficult to comprehensively capture the dynamic characteristics of the battery under complex working conditions. At the same time, it is difficult to balance between real-time performance and accuracy in existing methods, which easily leads to lagging risk assessment or misjudgment, especially under battery aging or extreme conditions. In addition, the lack of coordination between the vehicle side and the cloud side limits the depth of data analysis and the ability of personalized management.

[0003] These limitations reflect the core challenges in the field of power battery safety management. First, the complex correlation of battery state parameters is difficult to accurately model, and the interaction between electrochemical characteristics and external working conditions increases the prediction difficulty. Second, the accuracy of safety feature extraction and risk assessment is insufficient. Existing methods are prone to losing key information when processing multi-dimensional data, resulting in untimely and ineffective warning strategies. Finally, the adaptability problem of the model is prominent. Battery aging or working condition changes will reduce the prediction ability of the model, making it difficult to achieve long-term reliable online management. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes an online safety management and warning method for power batteries of electric vehicles to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides an online safety management and warning method for power batteries of electric vehicles, including:

[0006] Obtain multi-dimensional operation data of the power battery and perform preprocessing to obtain a standardized data set;

[0007] Based on the standardized data set, obtain dynamic working conditions and time series features, and generate a battery state feature set; if the battery state feature set does not meet the preset conditions, optimize the battery state feature set to obtain an optimized feature set;

[0008] Construct a risk assessment model, combine the optimized feature set to obtain the battery safety risk probability, and judge the risk level based on the battery safety risk probability;

[0009] If the risk level exceeds the threshold, a personalized management strategy generated by the cloud is obtained based on the optimized feature set and sent to the vehicle controller.

[0010] The online learning algorithm is used to fuse the real-time vehicle data and battery aging data, dynamically adjust the parameters of the risk assessment model, generate an adaptive model and output a graded warning signal.

[0011] Optionally, the process of obtaining the standardized data set includes:

[0012] Filtering technology is used to denoise the multi-dimensional operation data to generate a denoised data set; the abnormal data exceeding the preset threshold in the denoised data set is marked; the marked data is standardized to generate a standardized data set; among them, the multi-dimensional operation data includes voltage, current, temperature, and state of charge.

[0013] Optionally, the process of generating the battery state feature set includes:

[0014] Multi-dimensional features are obtained based on the standardized data set, and a convolutional neural network is used to extract the dynamic features under complex working conditions to obtain an initial feature set; a long short-term memory network is used to analyze the time series characteristics of the initial feature set to obtain a time series feature set; when the time series feature set does not match the preset working condition threshold, the time series feature set is re-clustered and the dynamic features and time series features are fused to generate a comprehensive feature set; according to the comprehensive feature set, a convolutional neural network is used to deeply analyze the feature set to extract the hidden working condition change features to obtain a deep feature set; if the correlation between the features in the deep feature set is lower than the preset threshold, the feature set is optimized by a dimensionality reduction technique to obtain an optimized feature set; through the optimized feature set, a long short-term memory network is used to model the long-term dependence relationship of the time series to generate a battery state feature set.

[0015] Optionally, the process of obtaining the optimized feature set includes:

[0016] If the voltage fluctuation in the battery state feature set exceeds the preset threshold, the feature set is smoothed by the Kalman filter algorithm to obtain an optimized feature set.

[0017] Optionally, the process of constructing a risk assessment model, obtaining the battery safety risk probability by combining the optimized feature set, and judging the risk level based on the battery safety risk probability includes:

[0018] Standardize the optimized feature set, extract the dynamic change features in the time series to obtain a dynamic feature set; if the signal fluctuation in the dynamic feature set exceeds the preset threshold, segment the data through the sliding window technique and calculate the statistical distribution characteristics to obtain a distribution feature set; use the random forest algorithm to train a classification model and output a risk probability value; when the risk probability value exceeds the threshold, calibrate the probability value through logistic regression and map it to a risk level.

[0019] Optionally, the process of obtaining a personalized management strategy generated by the cloud and sending it to the vehicle controller based on the optimized feature set includes:

[0020] Compress the optimized feature set through the vehicle controller to obtain a compressed feature set; encrypt and transmit the compressed feature set to the cloud through data encryption technology to obtain a cloud confirmation signal and get an upload confirmation status; if the upload confirmation status is successful, decrypt and analyze the encrypted feature set through a pre-trained deep learning model to obtain an analysis feature set; extract the dynamic change trend according to the analysis feature set to obtain a trend feature set; generate a personalized management strategy based on the trend feature set and determine the strategy execution parameters; send the strategy execution parameters to the vehicle controller through the cloud and obtain the execution confirmation status to judge whether the strategy deployment is completed.

[0021] Optionally, the process of dynamically adjusting the parameters of the risk assessment model through an online learning algorithm, fusing the vehicle-end real-time data and the battery aging data, generating an adaptive model and outputting a graded warning signal includes:

[0022] Obtain the battery operation data through the vehicle-end sensor and use the data fusion technology to integrate the battery operation data to obtain a fused data set; if the fused data set meets the preset conditions, use the online learning algorithm to update the parameters of the risk assessment model, adjust the model weights in combination with the battery aging data to obtain an adaptive model; extract the real-time dynamic characteristics from the multi-dimensional operation data according to the adaptive model, judge the safety state of the battery under complex working conditions, and output a warning signal.

[0023] Optionally, it further includes: obtaining the cloud historical operation data, combining the real-time dynamic characteristics, analyzing the battery aging trend by using a clustering algorithm to generate an aging feature set; updating the long-term prediction parameters of the adaptive model according to the aging feature set, predicting the risk probability under future working conditions, and outputting a dynamic management instruction.

[0024] The present invention also provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor executes the computer program to implement the steps of the above method.

[0025] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0026] Compared with the prior art, the present invention has the following advantages and technical effects:

[0027] The present invention discloses an online safety management and warning method for an electric vehicle power battery. By acquiring multi-dimensional operation data of the battery, data preprocessing and deep learning algorithms are used to extract dynamic features under complex working conditions, and a risk assessment model is constructed. When the risk level is relatively high, the present invention transmits the optimized feature set to the cloud, obtains personalized management strategies, and updates the vehicle-side model parameters using an online learning algorithm, and adjusts the weights in combination with battery aging data to achieve adaptive optimization. The present invention also uses a clustering algorithm to analyze the battery aging trend, updates long-term prediction parameters, and realizes dynamic prediction and management of the risk probability under future working conditions. This method can effectively evaluate the safety risks of power batteries under complex working conditions, provide real-time warnings and personalized management strategies, and improve the safety and lifespan of the batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation of this application. In the drawings:

[0029] Figure 1 is a flowchart of the method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and describe the application in detail with reference to the embodiments.

[0031] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0032] Embodiment 1

[0033] As Figure 1 shown, in this embodiment, an online safety management and warning method for an electric vehicle power battery is provided, including:

[0034] Acquire multi-dimensional operation data of the power battery and perform preprocessing to obtain a standardized data set;

[0035] As a specific implementation manner, the process of obtaining the standardized data set includes:

[0036] Filtering technology is used to denoise multi-dimensional operation data to generate a denoised data set; abnormal data exceeding a preset threshold in the denoised data set is marked; the marked data is normalized to generate a normalized data set; wherein, the multi-dimensional operation data includes voltage, current, temperature, and state of charge.

[0037] Specifically, multi-dimensional operation data is obtained from a power battery, including voltage data, current data, temperature data, and state of charge, and stored as an initial data set. Filtering technology is used to remove noise from the initial data set to generate a denoised data set. If the voltage data or current data in the denoised data set exceeds a preset threshold, the abnormal data is marked to obtain a marked data set. According to the marked data set, normalization technology is used to normalize the voltage data, current data, temperature data, and state of charge to generate a normalized data set.

[0038] Exemplarily, the acquisition of the initial data set needs to obtain voltage, current, temperature, and state of charge data from the power battery management system. Suppose an electric vehicle battery pack contains 100 single cells, and data is recorded once per second. The voltage range is 3.0 - 4.2V, the current is -50 to 200A, the temperature is -20 to 60°C, and the state of charge is 0 - 100%. These data are transmitted to the cloud in real time through sensors to form a high-frequency initial data set, laying a foundation for subsequent analysis. The high-frequency characteristics of the acquisition ensure the comprehensiveness of the data and help capture the subtle changes in battery operation.

[0039] Filtering technology is used to remove noise to generate a denoised data set. The commonly used median filtering can effectively smooth the abrupt noise in voltage and current. For example, for a certain battery voltage sequence of 3.6V, 3.61V, 3.8V, 3.62V, obviously 3.8V is an abnormal spike. Median filtering takes the median value 3.61V to replace the spike, eliminating the noise while retaining the trend. The denoised data set is smoother and can truly reflect the battery operation state, avoiding noise interference in subsequent analysis.

[0040] Abnormal data marking is aimed at the voltage and current in the denoised data set. The voltage threshold is set to 3.2 - 4.0V, and the current threshold is set to -30 to 180A. If the voltage is 4.1V at a certain moment, it is marked as overvoltage abnormal; if the current is 190A, it is marked as overcurrent abnormal. These marks form a marked data set, which is convenient for tracing the root cause of the problem. Abnormal marking improves the data quality and provides a warning basis for battery safety management.

[0041] The standardization technology normalizes four types of data in the labeled dataset. Since voltage, current, temperature, and state of charge have different dimensions, they need to be mapped to the 0-1 interval. For example, a voltage of 3.6V within the range of 3.2-4.0V is normalized to (3.6 - 3.2) / (4.0 - 3.2) = 0.5. The standardized dataset eliminates the influence of dimensions, enables different data to be comparable, and provides a consistent basis for feature extraction.

[0042] Based on the standardized dataset, dynamic operating conditions and time series features are obtained to generate a battery state feature set; if the battery state feature set does not meet the preset conditions, the battery state feature set is optimized to obtain an optimized feature set;

[0043] As a specific implementation manner, the process of generating the battery state feature set includes:

[0044] Based on the standardized dataset, multi-dimensional features are obtained, and a convolutional neural network is used to extract dynamic features under complex operating conditions to obtain an initial feature set; a long short-term memory network is used to analyze the time series characteristics of the initial feature set to obtain a time series feature set; when the time series feature set does not match the preset operating condition threshold, the time series feature set is re-clustered and the dynamic features and time series features are fused to generate a comprehensive feature set; according to the comprehensive feature set, a convolutional neural network is used to deeply analyze the feature set to extract hidden operating condition change features to obtain a deep feature set; if the correlation between the features in the deep feature set is lower than the preset threshold, the feature set is optimized through a dimensionality reduction technique to obtain an optimized feature set; through the optimized feature set, a long short-term memory network is used to model the long-term dependence relationship of the time series to generate a battery state feature set.

[0045] Exemplarily, when obtaining multi-dimensional features from the standardized dataset, it is necessary to ensure the diversity and consistency of the data. In the operating scenario of a power battery, the standardized dataset usually includes dimensions such as voltage, current, temperature, and state of charge. Suppose a battery pack of an electric vehicle collects data once per second, the voltage range is 3.0 - 4.2V, the current is -50 to 200A, the temperature is -20 to 60°C, and the state of charge is 0 - 100%. Key dimensions such as voltage and state of charge are extracted through principal component analysis to generate a multi-dimensional feature set containing the main information. This process retains the core characteristics of the data and lays a foundation for subsequent analysis.

[0046] The convolutional neural network is used to extract dynamic features under complex operating conditions. The convolutional neural network captures the local change patterns of voltage and current through convolutional kernels. For example, in an acceleration operating condition of a certain battery pack, the current rapidly rises from 50A to 150A, and the convolutional neural network identifies this mutation pattern and generates an initial feature set containing dynamic features such as the current change rate. This method can effectively capture the short-term fluctuations during operating condition switching.

[0047] The long short-term memory network analyzes the temporal characteristics of the initial feature set. The long short-term memory network is good at dealing with the long-term dependencies of time series. For example, in the case of a battery under high-temperature conditions for three consecutive days, the temperature remains at 45°C, and the state of charge gradually drops to 30%. Through the memory gate mechanism, the long short-term memory network extracts the temporal correlation between temperature and state of charge, forms a temporal feature set, and reflects the progressive change trend of the battery.

[0048] If the temporal feature set does not match the preset operating condition threshold, re-clustering is performed. For example, the preset high-temperature operating condition is that the temperature is higher than 40°C and the state of charge is lower than 50%, but a certain feature set shows that the temperature is 38°C and the state of charge is 20%. Through K-means clustering, the data is divided into high-temperature low-state-of-charge and other categories, and an adjusted feature set is generated. This adjustment ensures that the feature set is more in line with the actual operating conditions.

[0049] The feature fusion technology integrates dynamic features and temporal features. Using the weighted fusion method, dynamic features such as the current change rate account for 40% of the weight, and temporal features such as the temperature trend account for 60% of the weight. For example, when a battery is accelerating rapidly, the current change rate is high. After fusion, the comprehensive feature set reflects both instantaneous fluctuations and long-term trends, providing a comprehensive perspective for subsequent analysis.

[0050] The convolutional neural network conducts in-depth analysis on the comprehensive feature set to extract hidden operating condition change features. For example, when a battery starts at low temperature, the voltage drops abnormally fast. The convolutional neural network excavates the interaction pattern between voltage and temperature through deep convolution, generates a deep feature set, and reveals potential operating condition anomalies.

[0051] If the correlation of the deep feature set is lower than the threshold, such as the correlation coefficient is less than 0.7, it is optimized through dimensionality reduction technology. Principal component analysis can compress high-dimensional features such as voltage and current into the main components. For example, features that retain 80% of the variance are retained to generate an optimized feature set, reducing the subsequent calculation burden. For example, the long short-term memory network models the long-term dependencies of the optimized feature set as a time series. During the cyclic charge and discharge of a battery, the time required for the state of charge to drop from 80% to 20% gradually shortens. The long short-term memory network captures this trend and generates a battery state feature set, reflecting the capacity decay characteristics. This feature set provides an accurate basis for battery health management.

[0052] As a specific implementation manner, the process of obtaining the optimized feature set includes:

[0053] If the voltage fluctuation in the battery state feature set exceeds the preset threshold, the feature set is smoothed through the Kalman filter algorithm to obtain the optimized feature set;

[0054] Construct a risk assessment model, obtain the battery safety risk probability by combining the optimized feature set, and judge the risk level based on the battery safety risk probability. The process includes:

[0055] Standardize the optimized feature set, extract the dynamic change features in the time series to obtain a dynamic feature set; if the signal fluctuation in the dynamic feature set exceeds a preset threshold, segment the data through the sliding window technique and calculate the statistical distribution characteristics to obtain a distribution feature set; use the random forest algorithm to train a classification model and output a risk probability value; when the risk probability value exceeds the threshold, calibrate the probability value through logistic regression and map it to a risk level.

[0056] Specifically, according to the optimized feature set, preprocess the data using standardization techniques to obtain a standardized feature set. According to the standardized feature set, extract the dynamic change features in the time series to obtain a dynamic feature set. If the signal fluctuation in the dynamic feature set exceeds a preset threshold, segment the data through the sliding window technique to obtain a segmented feature set. According to the segmented feature set, calculate the statistical distribution characteristics of each segment feature to obtain a distribution feature set. According to the distribution feature set, use the random forest algorithm to train a classification model and output a risk probability value to obtain a probability feature set. If the risk probability in the probability feature set exceeds a preset threshold, calibrate the probability value through the logistic regression algorithm to obtain a calibrated feature set. According to the calibrated feature set, map the probability value to a risk level to determine the final risk level.

[0057] Exemplarily, assume that the voltage data range of a certain battery is from 3.2 to 4.2 volts and the current range is from 0.5 to 5 amperes. Through standardization, the voltage data may be converted to a range of -1.5 to 1.5, and the current data is processed similarly. This method can eliminate the influence of dimensions and facilitate subsequent analysis. According to the standardized feature set, extract the dynamic change features in the time series to obtain a dynamic feature set.

[0058] Specifically, the dynamic features may include the rate of change of voltage over time or the mutation points of current.

[0059] Suppose a certain battery pack is monitored, the rate of change of voltage is extracted, and it is observed that the voltage changes by about 0.01 volts per minute, but suddenly increases to 0.05 volts at a certain moment, which is recorded as a dynamic feature. This step focuses on capturing short-term fluctuations and trends in the data. If the signal fluctuation in the dynamic feature set exceeds a preset threshold, segment the data through the sliding window technique to obtain a segmented feature set.

[0060] The sliding window technique can divide the time series into segments of a fixed length. For example, with a 10-second window and a 2-second slide each time. Suppose the dynamic features of a certain battery show that the rate of change of voltage exceeds the threshold of 0.03 volts many times. Through the sliding window, the data of one day can be divided into hundreds of segments, and each segment contains the fluctuation information of voltage and current. This segmentation facilitates subsequent refined analysis. According to the segmented feature set, calculate the statistical distribution characteristics of each segment feature to obtain a distribution feature set.

[0061] Statistical distribution characteristics may include mean, variance, and skewness. For example, the mean voltage within a certain data segment is 3.7 volts, the variance is 0.02, and the skewness is 0.1, indicating that the data distribution is relatively concentrated and slightly skewed to the right. These statistical indicators reflect the regularity of each data segment and provide a basis for model training. According to the distribution feature set, a classification model is trained using the random forest algorithm, and a risk probability value is output to obtain a probability feature set.

[0062] The random forest determines the battery status risk through multiple decision trees. Assuming that a certain distribution feature is input, the model may output a risk probability of 0.75, indicating that this data segment corresponds to a high-risk state. The advantage of the random forest lies in its ability to model complex feature relationships. If the risk probability in the probability feature set exceeds a preset threshold, the probability value is calibrated through the logistic regression algorithm to obtain a calibrated feature set.

[0063] Logistic regression can fine-tune the probability output by the random forest. For example, assuming that the random forest outputs a risk probability of 0.78, exceeding the threshold of 0.7, logistic regression may calibrate it to 0.73 based on historical data to improve accuracy. This calibration can optimize the reliability of the probability value. According to the calibrated feature set, the probability value is mapped to a risk level to determine the final risk level.

[0064] The set probability of 0 to 0.3 is low risk, 0.3 to 0.6 is medium risk, and above 0.6 is high risk. Assuming that the calibrated probability is 0.73, it is mapped to high risk, indicating that the battery needs to be inspected. This mapping method is intuitive and easy to apply to actual monitoring scenarios.

[0065] If the risk level exceeds the threshold, based on the optimized feature set, a personalized management strategy generated by the cloud is obtained and sent to the vehicle controller;

[0066] As a specific implementation, the process of obtaining a personalized management strategy generated by the cloud and sending it to the vehicle controller based on the optimized feature set includes:

[0067] The vehicle controller compresses the optimized feature set to obtain a compressed feature set; the compressed feature set is encrypted and transmitted to the cloud through data encryption technology to obtain a cloud confirmation signal and an upload confirmation status; if the upload confirmation status is successful, the encrypted feature set is decrypted and analyzed through a pre-trained deep learning model to obtain an analysis feature set; according to the analysis feature set, the dynamic change trend is extracted to obtain a trend feature set; through the trend feature set, a personalized management strategy is generated, and the policy execution parameters are determined; the policy execution parameters are sent to the vehicle controller by the cloud and the execution confirmation status is obtained to determine whether the policy deployment is completed.

[0068] In a possible implementation, if the risk level exceeds a preset threshold, the vehicle controller compresses the optimized feature set to reduce the data transmission burden. The generation of the compressed feature set can be achieved through principal component analysis technology, which reduces the data dimension while retaining the main features.

[0069] For example, the original feature set contains more than a dozen dimensions such as battery temperature, voltage, and current. After compression, only three main components may be retained, and the data volume is reduced by about 70%, which is convenient for subsequent processing.

[0070] The compression process needs to ensure that key information is not lost to support the accuracy of subsequent analysis. Specifically, based on the compressed feature set, data encryption technology is used to protect data security, and the AES-256 encryption algorithm can be used to encrypt the data. For example, the compressed feature set is converted into an encrypted data packet, and the vehicle controller adds a unique identifier to each data packet before transmission to ensure that the data is not tampered with during transmission. This encryption method is particularly important in the vehicle networking scenario and can effectively prevent data leakage and improve system security.

[0071] The vehicle controller uploads the encrypted feature set to the cloud and waits for the cloud confirmation signal. For example, the controller sends the data packet to the cloud server through the 4G network, and the server returns a confirmation code marked as "upload successful" after receiving it. If the network is unstable, the controller can start the retransmission mechanism until it receives the confirmation signal. This mechanism ensures data integrity and provides a reliable basis for subsequent analysis.

[0072] After confirming the successful upload, the cloud uses a deep learning model to decrypt and analyze the encrypted feature set to generate an analysis feature set. A possible implementation is to use a convolutional neural network model to identify potential patterns in the data. For example, the model can analyze the change of battery voltage over time and detect abnormal fluctuations. The analysis feature set may include indicators such as the voltage mutation frequency, providing data support for subsequent trend extraction.

[0073] Based on the analysis feature set, the dynamic change trend is extracted to generate a trend feature set. Time series analysis technology can be used to identify the change law of battery performance over time. For example, the trend feature set may show that the temperature of a certain battery pack has been continuously rising in the past week, indicating a potential risk. This trend extraction helps to formulate more accurate countermeasures.

[0074] The trend feature set is used to generate a personalized management strategy and determine the strategy execution parameters. For example, for the trend of rising temperature, the strategy may include reducing the charging power or increasing the operation time of the cooling system. The parameters can be set to reduce the charging power to 80% and increase the operation frequency of the cooling system to once per hour. This strategy can effectively extend the battery life.

[0075] The policy execution parameters are sent from the cloud to the vehicle controller, and the execution confirmation status is obtained. For example, the cloud encapsulates the parameters into an instruction packet and sends it to the controller through an encrypted channel. After the controller executes, it returns the "deployment completed" status to ensure the implementation of the policy. It should be noted that the controller will verify the parameters before execution to avoid system exceptions caused by parameter errors.

[0076] The judgment of the execution confirmation status can be achieved through log records. For example, the controller generates execution logs, records the time and effect of policy execution, and uploads them to the cloud for subsequent auditing. This mechanism helps to track the policy effect and ensure the stability and reliability of the system operation.

[0077] The online learning algorithm is used to fuse the vehicle-end real-time data and battery aging data, dynamically adjust the parameters of the risk assessment model, generate an adaptive model, and output a graded warning signal.

[0078] As a specific implementation method, the process of using the online learning algorithm to fuse the vehicle-end real-time data and battery aging data, dynamically adjust the parameters of the risk assessment model, generate an adaptive model, and output a graded warning signal includes:

[0079] The battery operation data is obtained through vehicle-end sensors, and the data fusion technology is used to integrate the battery operation data to obtain a fused data set; if the fused data set meets the preset conditions, the online learning algorithm is used to update the parameters of the risk assessment model, and the model weights are adjusted in combination with the battery aging data to obtain an adaptive model; according to the adaptive model, the real-time dynamic characteristics are extracted from the multi-dimensional operation data, the safety status of the battery under complex working conditions is judged, and a warning signal is output.

[0080] Specifically, the battery operation data is obtained through vehicle-end sensors, and the data fusion technology is used to integrate the data to obtain a fused data set. If the fused data set meets the preset conditions, the online learning algorithm is used to update the parameters of the risk assessment model to obtain an updated parameter set. According to the updated parameter set, the weights of the risk assessment model are adjusted, and in combination with the battery aging data, an adaptive weight set is obtained. The adaptive weight set is used to optimize the risk assessment model to generate an optimized risk model. Through the optimized risk model, the fused data set is analyzed to judge the potential risk level and obtain risk level data. If the risk level data is higher than the preset threshold, a dynamic adjustment policy is generated through the vehicle controller to determine the policy parameters. According to the policy parameters, the vehicle control logic is updated to obtain an adjusted control state. According to the adaptive model, the real-time dynamic characteristics are extracted from the multi-dimensional operation data, the safety status of the battery under complex working conditions is judged, and a warning signal is output.

[0081] Further, multi-dimensional operation data is collected by sensors, and a clean data set is generated using data cleaning techniques. If the clean data set meets the preset conditions, the clean data set is processed by feature extraction techniques to obtain a dynamic feature set. Based on the dynamic feature set, a clustering algorithm is used to analyze the feature distribution to determine the working condition classification data. If the working condition classification data indicates a complex working condition, the dynamic feature set is processed by a state evaluation algorithm to obtain safety state data. Based on the safety state data, a threshold comparison technique is used to determine whether to trigger an alarm to obtain an alarm signal. Through the alarm signal, the control parameter set is updated to determine the adjusted control state.

[0082] Exemplarily, the sensors may include a temperature sensor, a voltage sensor, and a current sensor, which collect the operation state data of the battery pack in real time. For example, the temperature sensor records the surface temperature of the battery, the voltage sensor monitors the voltage of a single cell, and the current sensor captures the charge and discharge current. These data form a multi-dimensional data set, providing comprehensive information for subsequent analysis. The acquisition of multi-dimensional data needs to ensure time synchronization to ensure data consistency. Generating a clean data set using data cleaning techniques is a key step. Data cleaning aims to remove noise and outliers.

[0083] The temperature data can be processed by the moving window average method to remove the mutation values caused by sensor jitter. For example, in a certain acquisition, it is found that the temperature data jumps from 25 degrees to 50 degrees in a short time. The cleaning algorithm identifies it as an outlier and replaces it with the average value of the previous and subsequent data. In addition, missing values can be supplemented by linear interpolation. This method ensures the integrity and accuracy of the clean data set, laying a foundation for subsequent feature extraction. If the clean data set meets the preset conditions, a dynamic feature set is generated by feature extraction techniques.

[0084] The preset condition can be that the data integrity reaches more than 95%. The principal component analysis method can be selected as the feature extraction technique to reduce the multi-dimensional data to key features. For example, the cycle number and the internal resistance change rate are extracted from the voltage, current, and temperature data as dynamic features. These features reflect the operation state and potential degradation trend of the battery.

[0085] Feature extraction can also be combined with time series analysis to capture the variation law of data over time. Analyzing the feature distribution using a clustering algorithm to determine the working condition classification data is a process of further refinement analysis. The K-means clustering algorithm can be used to divide the dynamic feature set into normal working conditions, light load working conditions, and heavy load working conditions. For example, a battery pack with a low internal resistance change rate and stable temperature is classified as a normal working condition; while a rapid increase in internal resistance and high temperature are classified as heavy load working conditions. The clustering results help identify the complexity of the vehicle operation environment and provide a basis for state evaluation. If the working condition classification data indicates a complex working condition, the dynamic feature set is processed by a state evaluation algorithm to generate safety state data.

[0086] The state assessment algorithm can be based on fuzzy logic. By integrating features such as internal resistance, temperature, and number of cycles, it outputs a safety state score. For example, a score below 60 indicates a potential risk. Such a scoring mechanism intuitively reflects the battery health state and facilitates subsequent decision-making. Using threshold comparison technology to determine whether to trigger an early warning is the core of the decision-making. For example, when the safety state score is below 60 or the temperature exceeds 45 degrees, an early warning signal is triggered.

[0087] The early warning signal can be in a hierarchical form. A score of 50 to 60 is a low-level warning, and a score below 50 is a high-level warning. This hierarchical design facilitates the adoption of differentiated response measures. The final step is to update the control parameter set through the early warning signal and determine the adjusted control state.

[0088] The control parameter set includes the charging rate and the upper limit of power output. For example, when a high-level warning is triggered, the charging rate is reduced to 0.5C, and the upper limit of power output is reduced by 20%. Such adjustments can promptly respond to potential risks and ensure the stable operation of the vehicle.

[0089] As a specific implementation method, it also includes: obtaining historical operation data from the cloud, combining real-time dynamic characteristics, using a clustering algorithm to analyze the battery aging trend, generating an aging feature set; updating the long-term prediction parameters of the adaptive model according to the aging feature set, predicting the risk probability under future working conditions, and outputting a dynamic management instruction.

[0090] Specifically, obtain the battery historical operation data from the cloud, use data cleaning technology to process the historical operation data to obtain a cleaned data set. According to the cleaned data set, obtain the real-time collected dynamic characteristics, use feature fusion technology to process the cleaned data set and the dynamic characteristics to obtain a fused feature set. If the fused feature set meets the preset conditions, use a clustering algorithm to analyze the fused feature set to obtain aging trend data. According to the aging trend data, use feature extraction technology to process the aging trend data to obtain an aging feature set. If the aging feature set exceeds the range after comparison with the preset threshold, use the state assessment algorithm to process the aging feature set to obtain battery state data. According to the battery state data, use threshold comparison technology to judge the battery state data and determine the control parameter adjustment set. According to the control parameter adjustment set, update the operation control strategy and generate an adjusted control state.

[0091] Obtain feature data from the aging feature set, process the aging feature set using feature decomposition technology to obtain a decomposed feature set. If the decomposed feature set meets the preset conditions, then process the decomposed feature set using parameter adjustment technology to obtain updated model parameters. According to the updated model parameters, obtain future working condition data, and process the updated model parameters and future working condition data using probability prediction technology to obtain a risk probability distribution. If the risk probability distribution exceeds the preset threshold, then process the risk probability distribution using instruction generation technology to obtain a dynamic management instruction set. According to the dynamic management instruction set, process the dynamic management instruction set using instruction optimization technology to obtain an optimized instruction set. Through the optimized instruction set, update the operation control strategy to obtain an adjusted control state. According to the adjusted control state, process the adjusted control state using state feedback technology to obtain a feedback feature set.

[0092] This embodiment also provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor executes the computer program to implement the steps of the above method.

[0093] This embodiment also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the above method.

[0094] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An online safety management and warning method for power batteries of electric vehicles, characterized in that, It includes the following steps: Obtain multi-dimensional operation data of the power battery and perform preprocessing to obtain a standardized data set; Based on the standardized data set, obtain dynamic working conditions and time series features, and generate a battery state feature set; if the battery state feature set does not meet the preset conditions, optimize the battery state feature set to obtain an optimized feature set; Construct a risk assessment model, combine the optimized feature set to obtain the battery safety risk probability, and judge the risk level based on the battery safety risk probability; If the risk level exceeds the threshold, based on the optimized feature set, obtain a personalized management strategy generated by the cloud and send it to the vehicle controller; Through an online learning algorithm, fuse vehicle-end real-time data and battery aging data, dynamically adjust the parameters of the risk assessment model, generate an adaptive model and output a graded warning signal.

2. The online safety management and warning method for an electric vehicle power battery according to claim 1, wherein The process of obtaining the standardized data set includes: Use a filtering technique to denoise the multi-dimensional operation data to generate a denoised data set; mark the abnormal data in the denoised data set that exceeds the preset threshold; perform standardization processing on the marked data to generate a standardized data set; wherein, the multi-dimensional operation data includes voltage, current, temperature, and state of charge.

3. The online safety management and warning method for an electric vehicle power battery according to claim 1, wherein The process of generating the battery state feature set includes: Based on the standardized data set, obtain multi-dimensional features, use a convolutional neural network to extract dynamic features under complex working conditions to obtain an initial feature set; use a long short-term memory network to analyze the time series characteristics of the initial feature set to obtain a time series feature set; when the time series feature set does not match the preset working condition threshold, re-cluster the time series feature set and fuse dynamic features and time series features to generate a comprehensive feature set; according to the comprehensive feature set, use a convolutional neural network to deeply analyze the feature set, extract hidden working condition change features to obtain a deep feature set; if the correlation between the features in the deep feature set is lower than the preset threshold, optimize the feature set through a dimensionality reduction technique to obtain an optimized feature set; through the optimized feature set, use a long short-term memory network to model the long-term dependence relationship of the time series to generate a battery state feature set.

4. The online safety management and warning method for an electric vehicle power battery according to claim 2, wherein The process of obtaining the optimized feature set includes: If the voltage fluctuation in the battery state feature set exceeds the preset threshold, smooth the feature set through the Kalman filter algorithm to obtain an optimized feature set.

5. The online safety management and warning method for an electric vehicle power battery according to claim 1, wherein The process of constructing a risk assessment model, combining the optimized feature set to obtain the battery safety risk probability, and judging the risk level based on the battery safety risk probability includes: Standardize the optimized feature set, extract the dynamic change features in the time series to obtain a dynamic feature set; if the signal fluctuation in the dynamic feature set exceeds the preset threshold, segment the data by the sliding window technique and calculate the statistical distribution characteristics to obtain a distribution feature set; use the random forest algorithm to train a classification model and output a risk probability value; when the risk probability value exceeds the threshold, calibrate the probability value through logistic regression and map it to a risk level.

6. The online safety management and warning method for an electric vehicle power battery according to claim 1, characterized in that The process of obtaining a personalized management strategy generated by the cloud and sending it to the vehicle controller based on the optimized feature set includes: Compress the optimized feature set through the vehicle controller to obtain a compressed feature set; encrypt and transmit the compressed feature set to the cloud through data encryption technology to obtain a cloud confirmation signal and get an upload confirmation status; if the upload confirmation status is successful, decrypt and analyze the encrypted feature set through a pre-trained deep learning model to obtain an analysis feature set; extract the dynamic change trend according to the analysis feature set to obtain a trend feature set; generate a personalized management strategy through the trend feature set and determine the strategy execution parameters; send the strategy execution parameters to the vehicle controller through the cloud and obtain the execution confirmation status to judge whether the strategy deployment is completed.

7. The online safety management and warning method for an electric vehicle power battery according to claim 1, characterized in that The process of dynamically adjusting the parameters of the risk assessment model by fusing the vehicle-end real-time data and the battery aging data through an online learning algorithm, generating an adaptive model and outputting a graded warning signal includes: Obtain the battery operation data through the vehicle-end sensor and use the data fusion technology to integrate the battery operation data to obtain a fused data set; if the fused data set meets the preset conditions, use the online learning algorithm to update the parameters of the risk assessment model, adjust the model weights in combination with the battery aging data to obtain an adaptive model; extract the real-time dynamic characteristics from the multi-dimensional operation data according to the adaptive model, judge the safety state of the battery under complex working conditions and output a warning signal.

8. The online safety management and warning method for the power battery of an electric vehicle according to claim 1, characterized in that, It further includes: Obtain the cloud historical operation data, combine the real-time dynamic characteristics, and use the clustering algorithm to analyze the battery aging trend to generate an aging feature set; Update the long-term prediction parameters of the adaptive model according to the aging feature set, predict the risk probability under future working conditions, and output a dynamic management instruction.

9. A computer device, comprising: A memory and a processor with a computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the method according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1-8.

Images