Power battery data acquisition and storage method and device and medium

Abstract

The invention relates to the technical field of battery data processing. The power battery data acquisition and storage method comprises the following steps: acquiring operation parameter information of a power battery system in an operation process, constructing a multi-dimensional operation data sequence synchronous with time based on the operation parameter information, executing event trigger analysis on the multi-dimensional operation data sequence, and obtaining a data analysis result; if it is detected that any item in the operation parameter information exceeds a corresponding preset threshold value, high-frequency data sampling is started, high-frequency sampling information of a corresponding time period is obtained, change trend analysis is conducted on the high-frequency sampling information, and an evolution direction result of the abnormal operation state of the power battery system is obtained; and packaging the high-frequency sampling information and the evolution direction result to obtain a packaged data packet, and uploading the packaged data packet to a remote server. The method has the effect of solving the problem of insufficient sensing of the running state in the existing power battery system.

Description

Technical Field

[0001] The present invention relates to the technical field of battery data processing, and in particular to a method, device and medium for collecting and storing power battery data. Background Art

[0002] With the rapid development of electric vehicles, energy storage systems, and other fields, power batteries are becoming increasingly widely used as a key energy source. The performance and safety of power batteries are directly related to the operating efficiency and reliability of the entire system. Therefore, accurate monitoring and data analysis of power battery operating status have become crucial.

[0003] In existing technologies, the common practice for collecting power battery data is to regularly collect key battery parameters, such as voltage, current, and temperature, generally at fixed intervals. At the same time, some simple threshold alarm mechanisms are also set up to issue an alarm when a parameter exceeds the set normal range. In terms of data storage, local storage is often used, storing the collected data in the device's own storage media for subsequent review and analysis.

[0004] However, existing data collection and storage methods have significant flaws. Regular data collection methods struggle to capture sudden abnormalities during battery operation, and cannot obtain detailed abnormal data in a timely manner. Furthermore, simple threshold alarm mechanisms can only notify the occurrence of an anomaly but cannot predict or analyze its evolution. Locally stored data is susceptible to device failure or damage, leading to data loss. Summary of the Invention

[0005] In order to solve the problem of insufficient operating status perception in existing power battery systems, the present application provides a power battery data collection and storage method, device and medium.

[0006] The above-mentioned invention objective of this application is achieved through the following technical solutions: A power battery data collection and storage method, the power battery data collection and storage method comprising: Obtain operating parameter information of the power battery system during operation; Constructing a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; Performing event-triggered analysis on the multidimensional operation data sequence, and if it is detected that any item of the operation parameter information exceeds a corresponding preset threshold, initiating high-frequency data sampling to obtain high-frequency sampling information for the corresponding time period; Performing a change trend analysis on the high-frequency sampling information to obtain an evolution direction result of the abnormal operating state of the power battery system; The high-frequency sampling information and the evolution direction result are encapsulated to obtain an encapsulated data packet, and the encapsulated data packet is uploaded to a remote server.

[0007] The above technical solution addresses existing power battery system issues, such as the difficulty of low-frequency sampling in capturing transient anomalies, the high-frequency sampling process being limited by power supply conditions, anomaly detection limited to the threshold trigger level, and the lack of fault-tolerance mechanisms during data upload. Specifically, a time-synchronized, multi-dimensional operating data sequence constructed based on operating parameter information improves the temporal consistency and expressiveness of the operating data. Furthermore, high-frequency data sampling initiated through event triggering accurately locates the anomaly occurrence period and collects key details, addressing coverage gaps in fixed-period sampling. Furthermore, during the sampling process, power supply capacity is dynamically determined and wireless power switching strategies are implemented to ensure the continuity and stability of the high-frequency sampling process. Furthermore, a trained anomaly evolution trend recognition model is used to analyze the high-frequency sampling information, identifying the evolution direction and intensity of anomaly states, enhancing operational risk prediction capabilities. The analysis results are then packaged and uploaded with the high-frequency data. When transmission conditions are not met, the system automatically caches and appends time and status information to ensure data integrity and upload robustness, thereby achieving more accurate monitoring of the power battery system's operating status, stronger anomaly recognition capabilities, and a more reliable data storage mechanism.

[0008] In a preferred example, the present application may be further configured as follows: constructing a time-synchronized multi-dimensional operation data sequence based on the operation parameter information includes: Acquire data items containing time stamps in the operating parameter information, and perform interpolation alignment processing on the data items according to a unified time step to obtain an aligned operating parameter set; Based on a preset multidimensional data structure template, filling the information of each dimension in the aligned operating parameter set into a designated position of the preset multidimensional data structure template to form a multidimensional data format; The multidimensional data formats are spliced ​​and integrated in chronological order to obtain the multidimensional operation data sequence.

[0009] By adopting the above technical solution, by extracting data items containing time stamps and using a unified time step for interpolation and alignment processing, the timing misalignment problem caused by inconsistent acquisition frequencies of different parameters or data missing can be effectively solved, thereby generating a set of operating parameters with unified structure and time alignment. On this basis, based on the preset multi-dimensional data structure template, information of different dimensions is filled in the specified position in an orderly manner to ensure the accurate organization of various parameter data in the spatial dimension. Finally, a continuous multi-dimensional operating data sequence is formed through time sequence splicing and integration, so that subsequent event triggering analysis and trend modeling have a clear and consistent timing basis and data structure support, thereby providing data guarantee for the accurate identification and dynamic tracking of the operating status of the power battery system.

[0010] In a preferred example, the present application may be further configured as follows: based on a preset multidimensional data structure template, the information of each dimension in the aligned operating parameter set is correspondingly filled into a specified position of the preset multidimensional data structure template to form a multidimensional data format, including: Determine a corresponding time index based on the complete parameter content of each time step in the aligned operating parameter information set, and map the time index to the time axis dimension of the preset multidimensional data structure template to form a time distribution structure; Determine the corresponding position of the complete parameter content in the preset multidimensional data structure template according to the type and dimension of each parameter in the complete parameter content, and perform a filling operation to form a parameter distribution structure; According to the combination rule of the time distribution structure and the parameter mapping structure, the filling process of the preset multi-dimensional data structure template is completed to form the multi-dimensional data format.

[0011] By adopting the above technical solution, by analyzing the complete parameter content in each time step, a time index is generated based on the time mark corresponding to each time point, and the index is mapped to the time axis dimension in the preset multidimensional data structure template, ensuring that the data of different time nodes have a clear time positioning in the structure. Furthermore, by identifying the type attributes and functional dimensions of each parameter, its position coordinates in the multidimensional template are accurately determined, and data filling operations are performed to construct a structured data block with a complete parameter distribution relationship. Finally, by following the combination rules of the time distribution structure and the parameter mapping structure, the filling of the entire multidimensional template is systematically completed, so that the formed multidimensional data format has good consistency in terms of temporal continuity, structural integrity and dimensional correspondence.

[0012] In a preferred example, the present application may be further configured as follows: if it is detected that any item of the operating parameter information exceeds a corresponding preset threshold, high-frequency data sampling is initiated to obtain high-frequency sampling information of the corresponding time period, and further comprising: While executing the high-frequency data sampling, obtaining current power supply status information of the data recording path on which the high-frequency data sampling depends; Based on the power supply status information, it is determined whether the current power supply capacity meets the power condition required for the high-frequency data sampling; if the power supply status information does not meet the power condition required for the high-frequency data sampling, the power supply mode of the data recording path is switched to wireless power supply.

[0013] By adopting the above technical solution, it is possible to effectively deal with the problem of sampling interruption caused by insufficient power supply capacity during the high-frequency data sampling process. While executing high-frequency data sampling, the power supply status information of the data recording path related to the sampling process is obtained. By reading the power supply voltage, current or power supply stability parameters, it is determined whether the existing power supply meets the power condition threshold required for high-frequency sampling, such as whether it meets the minimum requirements for instantaneous power, voltage stability, etc. during continuous sampling. When the power supply status information indicates that the current power supply capacity is insufficient to support high-frequency data sampling, the power supply switching operation is triggered by a control instruction, and the original power supply path is switched to a wireless power supply mode, ensuring that high-frequency data sampling can continue when the main power supply is unstable or limited, thereby ensuring the integrity and continuity of the sampled data.

[0014] In a preferred example, the present application may be further configured as follows: performing a change trend analysis on the high-frequency sampling information to obtain an evolution direction result of the abnormal operating state of the power battery system includes: Arrange the high-frequency sampling information in the order of acquisition time to construct a high-frequency operation information sequence corresponding to the time; Inputting the high-frequency operation information sequence into the trained abnormal evolution trend recognition model to obtain trend feature information; Matching and analyzing the trend feature information with a preset abnormal evolution rule library to identify trend patterns related to the abnormal operating state of the power battery system; The evolution direction result is obtained according to the change direction and change intensity reflected by the trend pattern, combined with the corresponding time segment and parameter indicators.

[0015] By adopting the above technical solution, it is possible to achieve dynamic identification and evolution trend analysis of abnormal operating conditions of power battery systems. High-frequency sampling information is arranged in chronological order of collection time to construct a high-frequency operating information sequence that corresponds to time, ensuring continuity and traceability in the temporal dimension. This information sequence is then input into a trained abnormal evolution trend identification model. This model is constructed based on historical high-frequency operating characteristics and abnormal evolution labels and has the ability to recognize complex evolution patterns. After model calculation, trend information reflecting data change characteristics is extracted. This extracted trend feature information is then matched and analyzed with a preset abnormal evolution rule library. By comparing parameters such as similarity, key feature point distribution, and evolution direction, the abnormal trend pattern corresponding to the current high-frequency data sequence is identified. Finally, combined with the change direction and change intensity reflected by the trend pattern, the corresponding time segment is matched with the operating parameter indicators to obtain the evolution direction result reflecting the abnormal development path, providing accurate and real-time trend basis for subsequent status assessment and safety control.

[0016] In a preferred example, the present application may be further configured as follows: the trained abnormal evolution trend recognition model further includes: Acquire historical high-frequency sampling information of the power battery system and abnormal evolution label information corresponding to the historical high-frequency sampling information, and construct a supervised training sample set based on the historical high-frequency sampling information and the abnormal evolution label information; Inputting the supervised training sample set into the abnormal evolution trend recognition model to be trained to obtain a model verification result; Comparing the model verification result with a preset model performance evaluation index; if the model verification result does not meet the preset model performance evaluation index, updating the structure of the abnormal evolution trend recognition model to be trained, and continuing training until the model verification result meets the model performance evaluation index; When the model verification result meets the preset model performance evaluation index, the training of the abnormal evolution trend recognition model to be trained is stopped to obtain the trained abnormal evolution trend recognition model.

[0017] By adopting the above technical solution, it is possible to fully integrate historical high-frequency sampling information and corresponding abnormal evolution label information in the process of constructing an abnormal evolution trend identification model, and use labeled data to construct a supervised training sample set, so that the model has clear abnormal trend guidance during the training stage. The supervised training sample set is input into the model to be trained, and the model is initially trained and verified. The effectiveness of the model is judged based on the difference between the output results and the preset model performance evaluation indicators. If the model performance has not yet met the requirements, the model structure is dynamically updated and training is continued through structural optimization, parameter adjustment, etc., so as to improve the accuracy and generalization ability of the model in identifying abnormal trends. When the model verification results meet the performance indicator requirements, the training process is terminated to obtain a training completion model with stable structure and strong recognition ability, which provides a reliable model foundation for subsequent trend analysis of high-frequency operation data.

[0018] In a preferred example, the present application may be further configured as follows: encapsulating the high-frequency sampling information and the evolution direction result to obtain an encapsulated data packet, and uploading the encapsulated data packet to a remote server, further comprising: If, during the process of uploading the encapsulated data packet to the remote server, it is detected whether the current data transmission status meets the preset upload condition requirements, if the current data transmission status does not meet the preset upload condition requirements, the encapsulated data packet will be stored in the local buffer, and the corresponding time stamp and abnormal status information will be generated.

[0019] By adopting the above technical solution, it is possible to implement local temporary storage of encapsulated data packets when the conditions for remote data upload are not met, thereby avoiding direct data loss due to network anomalies, unstable signals or limited bandwidth. By saving the encapsulated data packets in the local buffer and additionally generating corresponding time stamps and abnormal status information, not only can the integrity and continuity of the data be ensured, but also the time and status basis for the subsequent retransmission mechanism can be provided, thereby enhancing the fault tolerance and stability of the data transmission process of the power battery system.

[0020] The second object of the present invention is achieved through the following technical solutions: A power battery data acquisition and storage device, comprising: An operating parameter acquisition module is used to obtain operating parameter information of the power battery system during operation; A data sequence construction module, configured to construct a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; An event trigger detection module is used to perform event trigger analysis on the multidimensional operation data sequence, and if it is detected that any item of the operation parameter information exceeds the corresponding preset threshold, high-frequency data sampling is initiated to obtain high-frequency sampling information of the corresponding time period; a trend analysis module, configured to perform a change trend analysis on the high-frequency sampling information to obtain an evolution direction result of the abnormal operating state of the power battery system; The encapsulation and uploading module is used to encapsulate the high-frequency sampling information and the evolution direction result to obtain an encapsulated data packet, and upload the encapsulated data packet to a remote server.

[0021] The above technical solution addresses existing power battery system issues, such as the difficulty of low-frequency sampling in capturing transient anomalies, the high-frequency sampling process being limited by power supply conditions, anomaly detection limited to the threshold trigger level, and the lack of fault-tolerance mechanisms during data upload. Specifically, a time-synchronized, multi-dimensional operating data sequence constructed based on operating parameter information improves the temporal consistency and expressiveness of the operating data. Furthermore, high-frequency data sampling initiated through event triggering accurately locates the anomaly occurrence period and collects key details, addressing coverage gaps in fixed-period sampling. Furthermore, during the sampling process, power supply capacity is dynamically determined and wireless power switching strategies are implemented to ensure the continuity and stability of the high-frequency sampling process. Furthermore, a trained anomaly evolution trend recognition model is used to analyze the high-frequency sampling information, identifying the evolution direction and intensity of anomaly states, enhancing operational risk prediction capabilities. The analysis results are then packaged and uploaded with the high-frequency data. When transmission conditions are not met, the system automatically caches and appends time and status information to ensure data integrity and upload robustness, thereby achieving more accurate monitoring of the power battery system's operating status, stronger anomaly recognition capabilities, and a more reliable data storage mechanism.

[0022] The third objective of this application is achieved through the following technical solutions: A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method when executing the computer program.

[0023] The fourth objective of this application is achieved through the following technical solutions: A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the above-mentioned method when executed by a processor.

[0024] In summary, this application includes at least one of the following beneficial technical effects: 1. It can address existing power battery system issues, such as the difficulty of low-frequency sampling in capturing transient anomalies, the high-frequency sampling process being limited by power supply conditions, anomaly detection limited to the threshold trigger level, and the lack of fault-tolerance mechanisms in the data upload process. Specifically, it constructs a time-synchronized, multi-dimensional operating data sequence based on operating parameter information, improving the temporal consistency and expressiveness of the operating data. Furthermore, high-frequency data sampling is initiated through event triggering, accurately locating the anomaly occurrence period and collecting key details, addressing the coverage gaps of fixed-period sampling. Furthermore, during the sampling process, it dynamically determines the power supply capacity and implements wireless power switching strategies to ensure the continuous stability of the high-frequency sampling process. Furthermore, it uses a trained anomaly evolution trend recognition model to perform trend analysis on the high-frequency sampling information, identifying the evolution direction and intensity of anomaly states, enhancing the ability to predict operational risks. Finally, the analysis results are packaged and uploaded with the high-frequency data. When transmission conditions are not met, the system automatically caches and appends time and status information to ensure data integrity and upload robustness, thereby achieving more accurate monitoring of the power battery system's operating status, stronger anomaly recognition capabilities, and a more reliable data storage mechanism. 2. It is able to fully integrate historical high-frequency sampling information and corresponding abnormal evolution label information in the process of building an abnormal evolution trend identification model, and use labeled data to construct a supervised training sample set, so that the model has clear abnormal trend guidance during the training stage. The supervised training sample set is input into the model to be trained for preliminary model training and verification. The effectiveness of the model is judged based on the difference between the output results and the preset model performance evaluation indicators. If the model performance has not yet met the requirements, the model structure is dynamically updated and training is continued through structural optimization, parameter adjustment, etc., so as to improve the accuracy and generalization ability of the model in identifying abnormal trends. When the model verification results meet the performance indicator requirements, the training process is terminated, and a training model with stable structure and strong recognition ability is obtained, providing a reliable model foundation for subsequent trend analysis of high-frequency operation data; 3. It can realize local temporary storage processing of encapsulated data packets when the remote data upload conditions are not met, avoiding direct data loss due to network anomalies, unstable signals or limited bandwidth. By saving the encapsulated data packets in the local buffer and additionally generating corresponding time stamps and abnormal status information, it can not only ensure the integrity and continuity of the data, but also provide time and status basis for the subsequent retransmission mechanism, thereby enhancing the fault tolerance and stability of the data transmission process of the power battery system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flow chart of a method for collecting and storing power battery data in one embodiment of the present application; Figure 2This is a flowchart for implementing step S20 in a method for collecting and storing power battery data in one embodiment of the present application; Figure 3 This is a flowchart for implementing step S202 in a method for collecting and storing power battery data in one embodiment of the present application; Figure 4 This is another implementation flowchart of step S30 in a method for collecting and storing power battery data in one embodiment of the present application; Figure 5 This is a flowchart for implementing step S40 in a method for collecting and storing power battery data in one embodiment of the present application; Figure 6 This is a flowchart for implementing step S402 in a method for collecting and storing power battery data in one embodiment of the present application; Figure 7 This is a flowchart for implementing step S50 in a method for collecting and storing power battery data in one embodiment of the present application; Figure 8 This is a principle block diagram of a power battery data acquisition and storage device in one embodiment of the present application; Figure 9 It is a schematic diagram of a device in one embodiment of the present application. DETAILED DESCRIPTION

[0026] The present application is further described in detail below with reference to the accompanying drawings.

[0027] In one embodiment, if Figure 1 As shown, the present application discloses a method for collecting and storing power battery data, which specifically includes the following steps: S10: Obtaining operating parameter information of the power battery system during operation.

[0028] Operating parameter information refers to the key indicator data reflecting the system status and performance collected during the operation of the power battery system. The data includes but is not limited to battery cell voltage, total voltage, current, temperature, charge and discharge status, SOC (state of charge), SOH (state of health), number of cycles and environmental parameters. The parameters have clear timestamps and can reflect the dynamic evolution of the system status.

[0029] Specifically, during the operation of the power battery, the voltage sensor connected to the voltage sampling loop, the current sensor in the current sampling loop, and the temperature sensing device arranged on the surface of the battery cell are used to periodically sample the operating conditions such as the single cell voltage, the whole pack voltage, the charge and discharge current, the cell temperature, and the ambient temperature. During the acquisition process, various analog signals are converted into standard input signals after being processed through voltage conditioning and anti-interference filtering. Subsequently, the processed analog signals are numerically quantized using an analog-to-digital converter, and the corresponding physical quantity units are uniformly assigned according to the parameter types. Data record items containing time tags are generated based on the sampling time, and all parameter record items collected at each time point are combined into a set of operating parameter information.

[0030] S20: Constructing a multi-dimensional operation data sequence synchronized with time based on the operation parameter information.

[0031] Specifically, data record items containing time tags are extracted from the collected operating parameter information, and linear interpolation operations are performed on all record items according to a unified time step. By constructing a continuous function model between the sampled time points, the values ​​of the interpolation points are calculated to fill in the missing data, ensuring that all parameters have complete values ​​at each time step, and thus forming an aligned operating parameter set with a consistent time reference; subsequently, according to a preset multidimensional data structure template, a multidimensional information structure including a time axis, a voltage axis, a current axis, a temperature axis and other extended parameter axes is established, in which the aligned dimensional parameters are respectively filled into the corresponding dimensional slots, and the binding operation between each data point on the time axis and each parameter dimension is completed through the index mapping function, so that the operating information of all dimensions at that moment can be retrieved at any time point, and finally the filled data units are arranged in time order to construct a multidimensional operating data sequence with a unified time reference and parameter structure.

[0032] S30: Execute event trigger analysis on the multi-dimensional operation data sequence. If it is detected that any item in the operation parameter information exceeds the corresponding preset threshold, high-frequency data sampling is started to obtain high-frequency sampling information of the corresponding time period.

[0033] In this embodiment, the preset threshold refers to the boundary value of the normal operating range defined for each indicator in the operating parameter information. This boundary value can be set based on historical experience data, statistical analysis results, or battery manufacturer recommendations. It is used to determine whether the parameter exceeds the normal range and trigger corresponding subsequent actions. High-frequency data sampling refers to the process of rapidly and continuously collecting data on the relevant operating parameters at a higher frequency than the normal sampling rate when an abnormal operating parameter is detected. The sampling interval can be shortened to milliseconds to capture transient characteristics during the occurrence and evolution of abnormal events.

[0034] Specifically, the constructed multi-dimensional operation data sequence is used as the input information source. For each time step in the data sequence, the current values ​​of key operating parameters such as voltage, current, and temperature are read item by item, and the pre-set parameter threshold configuration table is called. The current value is judged whether it exceeds the corresponding safe operating range by comparing item by item. During the detection process, in order to avoid misjudgment caused by instantaneous fluctuations, a sliding window mechanism is introduced to calculate the average value and fluctuation range of the parameter in a short period of time as the basis for judgment. If a parameter value in any window continuously exceeds the set upper limit or falls below the set lower limit, it is determined that an abnormal trigger event has occurred. Subsequently, the high-frequency sampling starting position is marked according to the current trigger time point, and the sampling frequency control logic is switched. The high-frequency sampling process is started on the basis of the original low-frequency sampling, and the operation data with higher time resolution within this period is continuously obtained to capture potential transient abnormal information. The collected result is the high-frequency sampling information corresponding to the event.

[0035] S40: Analyze the change trend of the high-frequency sampling information to obtain the evolution direction result of the abnormal operating state of the power battery system.

[0036] Specifically, the acquired high-frequency sampling information is reordered in chronological order to ensure the continuity and comparability of each parameter between adjacent sampling points. The difference and slope of key parameters such as voltage, current and temperature between consecutive sampling points are calculated point by point in a traversal manner to form a differential sequence describing the parameter change rate and amplitude. At the same time, local change fragments are extracted based on the set time window length, and their average slope, change extreme value and fluctuation period are calculated to construct the corresponding change trend feature vector. The change trend feature vector is used as input data and input into the trained abnormal evolution trend recognition model. The model is obtained based on historical abnormal case learning and is used to identify the evolution pattern corresponding to the current trend. In the model, convolution operation is performed to extract key trend features, and the corresponding trend label and evolution direction indicator are output in combination with the classification structure. Finally, based on the evolution path type, time gradient and parameter change direction indicated by the trend label and direction indicator, an evolution direction result is generated to characterize the future evolution trend of the current abnormal state.

[0037] S50: Encapsulate the high-frequency sampling information and the evolution direction result to obtain an encapsulated data packet, and upload the encapsulated data packet to a remote server.

[0038] Specifically, the high-frequency sampling information and the corresponding evolution direction results are formatted according to a unified data structure specification. First, the key parameter fields and corresponding timestamps in the high-frequency sampling information are extracted and arranged in chronological order. Then, an association is established between each sampling record and its corresponding trend analysis result. The trend type, change amplitude, change rate and other indicators in the evolution direction results are filled into the corresponding positions in the preset data structure template through field mapping. After the field filling is completed, a verification operation is performed to ensure that each data field meets the type and length specifications. After the verification passes, the formatted data structure is encapsulated in a compression encoding method to generate a data packet, and the construction time information, sampling window number and unique identification number are attached to the data packet to support subsequent data tracing and indexing. Finally, the upload channel of the remote server is determined according to the current network status, and the data communication instruction is called to send the encapsulated data packet to the remote server through the specified protocol and complete the confirmation receipt.

[0039] In one embodiment, if Figure 2 As shown, in step S20, a multi-dimensional operation data sequence synchronized with time is constructed based on the operation parameter information, including: S201: Acquire data items containing time stamps in the operating parameter information, perform interpolation and alignment processing on the data items according to a unified time step, and obtain an aligned operating parameter set.

[0040] Specifically, various raw data items with time stamps are extracted from the operating parameter information, including parameters such as voltage, current, temperature, and SOC. The sampling timestamps corresponding to each type of parameter are extracted to form a time series. Subsequently, the minimum sampling interval appearing in all parameters is counted as a unified time step reference to construct a target alignment time axis. At each target time point, it is determined whether each parameter has a valid value. If there is a missing value, a linear interpolation method is used to fill it in. The interpolation alignment process is as follows: suppose two adjacent known time points are t1 and t2, and the corresponding parameter values ​​are v1 and v2. If the target time point is t, and t1 < t < t2 is satisfied, then the interpolation value is v = v1 + (v2 - v1) × (t - t1) / (t2 - t1). After completing the interpolation filling for all parameters in this way, the parameter values ​​are combined in order according to the target time axis to form an aligned operating parameter set under a unified time reference.

[0041] S202: Based on a preset multidimensional data structure template, fill the information of each dimension in the aligned operating parameter set into a corresponding designated position of the preset multidimensional data structure template to form a multidimensional data format.

[0042] In this embodiment, the preset multidimensional data structure template refers to a structured data organization model constructed based on the classification dimension, time dimension and logical grouping relationship of the power battery system operating parameters. It is usually expressed in the form of a tensor or a nested array structure, which is used to unify the time axis and dimension correspondence of different parameters to facilitate subsequent data modeling and analysis processing.

[0043] Specifically, first, according to the parameter types and time series lengths contained in the aligned operating parameter set, an established preset multidimensional data structure template is selected. The template has a definition structure of a time axis dimension and multiple parameter dimensions. For example, rows represent different time steps, and columns represent various operating parameters, such as single-cell voltage, temperature, current, etc. For each time step, the complete parameter set corresponding to the time step is traversed, and its filling position in the matrix is ​​determined according to the mapping rule of the parameter in the template. The mapping rule predefines the correspondence between the parameter name and the template coordinate. During the filling process, the corresponding parameter values ​​are filled one by one into the corresponding cells of the template structure according to the time index of each time step, ensuring that the same parameter is filled to the same column coordinate in all time steps, while ensuring the orderly arrangement of data in the time dimension, and finally converting all aligned operating parameter sets into a multidimensional data format that meets the requirements of a unified time base and parameter dimension definition.

[0044] S203: The multidimensional data formats are spliced ​​and integrated in chronological order to obtain a multidimensional running data sequence. Specifically, the filled multidimensional data format is sorted according to the time index to ensure that each data record is strictly increasing in the time dimension. Then, the multidimensional data formats generated in each time period are connected in sequence according to the preset splicing rules. The splicing rules stipulate the seamless connection method between adjacent data segments to maintain time continuity and parameter dimension consistency. During the splicing process, the possible time overlap or gaps between adjacent paragraphs are corrected. The overlapping parts retain the latest data according to the time priority, and the gaps are filled by interpolation or null value marking to maintain the integrity of the sequence. After the splicing and integration are completed, a multidimensional operation data sequence covering continuous time intervals, consistent parameter dimensions and complete structure is obtained.

[0045] In one embodiment, if Figure 3 As shown, in step S202, based on the preset multidimensional data structure template, the information of each dimension in the aligned operating parameter set is filled into the specified position of the preset multidimensional data structure template to form a multidimensional data format, including: S2021: Determine the corresponding time index based on the complete parameter content of each time step in the aligned operating parameter information set, and map the time index to the time axis dimension of the preset multidimensional data structure template to form a time distribution structure.

[0046] Specifically, each record in the aligned operating parameter information set is traversed, and the time stamp in each record is extracted as a time index. By setting a unified time step and starting reference time, the time index is mapped to a time series number in the form of an integer, and the specific position on the time axis dimension in the multidimensional data structure template is located accordingly; the linear mapping rule is adopted in the mapping process to ensure that the distribution of each time step in the template maintains a consistent interval. When a time step cannot be accurately mapped to the template position, the nearest neighbor strategy is used for mapping compensation, and finally an accurate association between the time index and the actual data record is established in the multidimensional data structure template, forming a time distribution structure with a strict time progressive relationship.

[0047] S2022: Determine the corresponding position of the complete parameter content in the preset multidimensional data structure template according to the type and dimension of each parameter in the complete parameter content, and perform a filling operation to form a parameter distribution structure.

[0048] Specifically, the complete parameter content corresponding to each time step is parsed to identify the physical meaning and data type of each operating parameter, including voltage, current, temperature, SOC, etc., and these parameters are classified into the corresponding physical dimension, statistical dimension or functional dimension in the multidimensional data structure template according to the preset parameter dimension classification rules; when performing the filling operation, according to the mapping relationship between the parameter classification and the dimension label in the template structure, the parameter values ​​are filled in the specified position of the corresponding dimension one by one. For non-existent parameter dimensions, empty values ​​are retained or filled with default placeholders to ensure the consistency and integrity of the template structure. Finally, a clearly defined parameter distribution structure is formed in the template, reflecting the organizational position of different types of parameters in the multidimensional space.

[0049] S2023: According to the combination rule of the time distribution structure and the parameter mapping structure, the filling process of the preset multidimensional data structure template is completed to form a multidimensional data format.

[0050] In this embodiment, the combination rules of the parameter mapping structure refer to the definition method of the corresponding filling positions of various types of aligned operating parameter information in the multidimensional data structure template. The rules include the mapping relationship between the time index and the parameter dimension, the mapping relationship between the parameter type and the structural position, and the sequence and coverage logic followed in the filling process.

[0051] Specifically, combining each time index in the time distribution structure with the filling position of the corresponding dimension in the parameter distribution structure, the template filling process is executed according to the combination rules of the parameter mapping structure. The combination rules of the parameter mapping structure include a dual-axis mapping method with time as the first dimension and the parameter dimension as the second dimension, and accurately corresponding the complete parameter set of each time step to the row and column coordinates of the multidimensional data structure template. During the filling process, unified format conversion and numerical standardization are performed on the data of each time step to ensure the consistency and comparability of the data format. If there is missing data, it is interpolated or marked as missing according to the set completion strategy, and finally the effective area of ​​the entire multidimensional data structure template is filled to form a multidimensional data format with time series characteristics and parameter dimension correlation.

[0052] In one embodiment, if Figure 4 As shown, in step S30, if it is detected that any item of the operating parameter information exceeds the corresponding preset threshold, high-frequency data sampling is started to obtain high-frequency sampling information of the corresponding time period, which also includes: S301: While executing high-frequency data sampling, obtain current power supply status information of a data recording path on which the high-frequency data sampling depends.

[0053] Specifically, while starting high-frequency data sampling, the power supply status information on the data recording path through which the high-frequency sampling operation is currently performed is obtained by reading real-time electrical parameters such as voltage, current and power factor in the power supply line. The power supply status information includes the type of power supply, power supply output stability, voltage fluctuation range and instantaneous power output capability. During the acquisition process, the power supply status can be continuously detected in combination with the set sampling period, and the status data at each detection time point can be recorded to reflect the dynamic changes of the power supply status over time, providing basic information support for the subsequent judgment of whether the high-frequency sampling power conditions are met.

[0054] S302: Based on the power supply status information, determine whether the current power supply capacity meets the power conditions required for high-frequency data sampling. If the power supply status information does not meet the power conditions required for high-frequency data sampling, switch the power supply mode of the data recording path to wireless power supply.

[0055] In this embodiment, the power condition refers to the minimum power supply threshold required to support high-frequency data sampling operations. The power threshold is determined based on technical parameters such as the operating current, voltage requirements, and sampling frequency of the sampling circuit. If the current power supply capacity is lower than the threshold, sampling interruption or data loss may occur.

[0056] Specifically, in the process of judging whether the current power supply capacity meets the power conditions required for high-frequency data sampling based on the power supply status information, the power requirement standard corresponding to the high-frequency data sampling operation is first found to determine the required minimum operating power threshold and the power supply voltage stability range. Then, the instantaneous power supply power value is calculated based on the real-time detected voltage and current data, and it is evaluated whether it is in the stable working range. If it is found that the current power supply power is lower than the minimum power threshold, or the power supply voltage fluctuation exceeds the allowable range, it is determined that the current power supply capacity cannot meet the power requirement of high-frequency data sampling. Subsequently, according to the set power supply switching strategy, the power management component is controlled to shut down the original power supply path, and the wireless power supply unit is activated to establish a stable wireless power supply link, thereby ensuring the continuous power supply required for sampling without interrupting the sampling process.

[0057] Furthermore, the power supply switching strategy refers to a decision-making mechanism that automatically switches from wired power supply to wireless power supply according to the set logic when the current power supply capacity is insufficient to support high-frequency data sampling tasks. This strategy comprehensively considers factors such as power consumption, remaining power of the device, and reliability of wireless power supply to ensure the stability and continuity of the sampling process.

[0058] In one embodiment, if Figure 5 As shown, in step S40, the high-frequency sampling information is subjected to a change trend analysis to obtain the evolution direction of the abnormal operating state of the power battery system, including: S401: Arrange the high-frequency sampling information in the order of acquisition time to construct a high-frequency operation information sequence corresponding to the time.

[0059] Specifically, in the process of arranging the high-frequency sampling information in order of acquisition time, all high-frequency sampling data records obtained within the current trigger event time window are first extracted from the data buffer, the timestamp information attached to each record is extracted, and the timestamp is used as the sorting basis. A stable time sorting algorithm is used to sort all the sampling data in ascending order. After the sorting is completed, the sorted data is structured and stored in sequence, and a mapping relationship is constructed between each time point and a complete set of operating parameter items, further forming a time series structure. In this structure, each time node has corresponding multi-dimensional operating information such as voltage, current, and temperature, thereby forming a high-frequency operating information sequence with time consistency and parameter integrity.

[0060] S402: Inputting the high-frequency operation information sequence into the trained abnormal evolution trend recognition model to obtain trend feature information.

[0061] In this embodiment, the trained abnormal evolution trend recognition model refers to a converged machine learning model obtained by training with labeled high-frequency sampling historical data in a supervised learning manner.

[0062] Specifically, when the high-frequency operation information sequence is input into the trained abnormal evolution trend recognition model, the high-frequency operation information sequence is first formatted according to the input requirements of the model, including normalizing the value of each dimension parameter to the standard input range supported by the model, filling in missing fields or removing abnormal data points, ensuring that the input data structure is consistent with the model training stage, and then constructing the processed multidimensional data into continuous sliding window samples in chronological order. Each window contains operation information of several consecutive moments, and serves as a one-time input of the model. These sliding window samples are input into the recognition model in turn. The model automatically extracts the implicit features representing the change trend of the operation status in each window based on its internally constructed feature extraction unit, such as the temporal neural network structure or the convolution structure, and outputs the intermediate feature vector reflecting the characteristics of the abnormal evolution status. Finally, all the intermediate feature vectors are summarized to obtain the trend feature information.

[0063] S403: Matching and analyzing the trend feature information with a preset abnormal evolution rule library to identify trend patterns related to abnormal operating states of the power battery system.

[0064] Specifically, when matching and analyzing trend feature information with the preset abnormal evolution rule library, first determine the abnormal feature dimension category to which it belongs based on the parameter name, unit and change direction corresponding to each characteristic value in the trend feature information, such as parameter dimensions such as voltage drop gradient, temperature rise rate, and internal resistance mutation amplitude. Then, based on the characteristic value range, change amplitude and time distribution of each parameter dimension, map the trend feature information one by one to the rule entries defined in the abnormal evolution rule library. Each rule entry contains a complete dimension structure, parameter threshold range and corresponding abnormal identifier. After the mapping is completed, all candidate rule entries are ranked by comprehensively scoring by calculating the parameter matching degree, time overlap rate and change trend consistency score between the trend feature information and each rule entry, and select the abnormal pattern identifier with the highest score as the trend pattern corresponding to the current abnormal operating state of the power battery system.

[0065] S404: Obtain an evolution direction result based on the change direction and change intensity reflected by the trend pattern, combined with the corresponding time segment and parameter indicators.

[0066] Specifically, according to the change direction and intensity reflected by the trend pattern, combined with the corresponding time segment and parameter indicators, first extract all the constituent elements from the trend pattern, including the types of operating parameters involved, the change trend of each parameter (such as continuous rise, periodic fluctuation or sharp decline) and a quantitative description of the change amplitude. Then, combined with the time segment covered by the trend pattern, determine the evolution path of each parameter within the segment. Next, arrange the parameter evolution behavior in the evolution path in sequence according to the time axis, identify the key change nodes and extreme points of the change rate, and construct a parameter change vector set based on the actual numerical values ​​in the parameter indicators. The vector set is used to characterize the joint change direction and intensity of each parameter within the time segment. Finally, by comprehensively analyzing the direction consistency and intensity distribution of the vector set, the reflected abnormal evolution direction result is output to represent the abnormal trend characteristics of the current power battery system operation status.

[0067] In one embodiment, if Figure 6 As shown, in step S402, the abnormal evolution trend recognition model that has been trained also includes: S4021: Obtain historical high-frequency sampling information of the power battery system and abnormal evolution label information corresponding to the historical high-frequency sampling information, and construct a supervised training sample set based on the historical high-frequency sampling information and the abnormal evolution label information.

[0068] Specifically, the historical high-frequency sampling information recorded by the power battery system in multiple historical operating cycles is obtained. The historical high-frequency sampling information includes the continuously changing values ​​of multi-dimensional operating parameters such as voltage, current, temperature, and internal resistance at each time point. By traversing the high-frequency sampling segments in each historical cycle, the time interval related to the known abnormal state is extracted, and the abnormal evolution label information generated by manual annotation or the abnormality detection system in the time interval is synchronously retrieved. The abnormal evolution label information includes the abnormal type, evolution stage and its corresponding time range; then the high-frequency sampling information in each historical time period is used as the sample input part, and its corresponding abnormal evolution label information is used as the supervision output part. Sample pairs are constructed according to a fixed time window length and a sliding step size. After obtaining multiple sample pairs, they are uniformly converted into a format that adapts to the input requirements of the model, and finally a supervised training sample set containing the correspondence between the input sequence and the output label is formed.

[0069] S4022: Input the supervised training sample set into the abnormal evolution trend recognition model to be trained to obtain the model verification result.

[0070] Specifically, each group of samples in the supervised training sample set is input into the abnormal evolution trend recognition model to be trained in turn. The model is a deep learning model with time series modeling capabilities, and adopts a bidirectional long short-term memory network structure to extract temporal correlation features in the input sequence; during the training process, the high-frequency sampling sequence in the sample set is used as input, and the abnormal evolution label is used as the target output. The cross entropy loss or mean square error loss between the model prediction result and the label is calculated, and the model parameters are updated through the back propagation algorithm; after completing a round of training, the model is verified using the reserved validation sample subset, and the model's predicted output on the subset is compared with the label information. The model performance is comprehensively evaluated based on multiple evaluation indicators such as precision, recall rate, and F1 score to obtain the model verification result of the current training round.

[0071] S4023: Compare the model verification result with the preset model performance evaluation index. If the model verification result does not meet the preset model performance evaluation index, update the structure of the abnormal evolution trend recognition model to be trained and continue training until the model verification result meets the model performance evaluation index.

[0072] In this embodiment, the preset model performance evaluation indicators refer to quantitative evaluation criteria used to measure the quality of the model during the training process, including but not limited to accuracy, recall rate, F1 score, loss function value, etc. The model is considered to be trained after reaching the set threshold on the validation set.

[0073] Specifically, based on the numerical performance of various performance indicators in the model verification results, including but not limited to accuracy, recall rate, precision and F1 score, they are compared item by item with the corresponding preset performance evaluation thresholds. When any indicator fails to meet the corresponding threshold requirements, the current structure of the model may have insufficient or overfitting problems, and the model is adjusted using a structural update strategy, such as increasing the number of network layers, expanding the number of neurons, introducing a regularization mechanism or replacing the activation function. After the structural adjustment, the updated model parameters are reinitialized, and the supervised training sample set is re-input into the updated model for continued training. By iteratively executing the model structure update and training process, the model performance indicators gradually approach the preset standards until all performance indicators meet the corresponding evaluation requirements. At this time, the training of the abnormal evolution trend recognition model to be trained is considered to be completed.

[0074] S4024: When the model verification result meets the preset model performance evaluation index, the training of the abnormal evolution trend recognition model to be trained is stopped to obtain a trained abnormal evolution trend recognition model.

[0075] Specifically, after completing each round of model training, the performance of the current model on the validation set is counted, and multiple performance indicators including accuracy, recall rate, precision, F1 score, etc. are extracted and compared one by one with the corresponding preset model performance evaluation indicators. When all performance indicators reach or exceed the set evaluation threshold, it is judged that the current model has a stable abnormal trend recognition capability, the subsequent training process is terminated, and the model structure and its corresponding parameter status after the current round of training are saved as the final trained abnormal evolution trend recognition model.

[0076] In one embodiment, if Figure 7 As shown, in step S50, the high-frequency sampling information and the evolution direction result are encapsulated to obtain an encapsulated data packet, and the encapsulated data packet is uploaded to a remote server, which also includes: S501: If, during the process of uploading the encapsulated data packet to the remote server, it is detected whether the current data transmission status meets the preset upload condition requirements, if the current data transmission status does not meet the preset upload condition requirements, the encapsulated data packet will be stored in the local buffer, and the corresponding time stamp and abnormal status information will be generated.

[0077] In this embodiment, the preset upload condition requirements refer to the minimum communication quality standards that must be met for data upload to the remote server. The standards may include threshold settings for indicators such as network bandwidth, delay, packet loss rate, and signal strength, which are used to determine whether the current network status is suitable for data upload operations.

[0078] Specifically, when performing an upload operation, the real-time data transmission performance index is calculated based on the bandwidth, delay, packet loss rate and other parameters of the current communication link, and compared with the performance threshold in the preset upload condition requirements. When any index is lower than the set threshold, it is determined that the upload condition requirements are not met. At this time, the data packet upload process is terminated, and the encapsulated data packet to be uploaded is written into the device's built-in local buffer storage unit. At the same time, the current system time is obtained to generate a corresponding time stamp, and the abnormal status information is generated in combination with the identification field of the abnormal operation status. The time stamp and the abnormal status information are attached to the metadata field of the encapsulated data packet to facilitate data retransmission and abnormality tracking when the network is restored.

[0079] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0080] In one embodiment, a power battery data acquisition and storage device is provided, which corresponds one-to-one to a power battery data acquisition and storage method in the above embodiment. Figure 8As shown, the power battery data acquisition and storage device includes an operating parameter acquisition module, a data sequence construction module, an event trigger detection module, a trend analysis module, and a packaging and upload module. The functional modules are described in detail as follows: An operating parameter acquisition module is used to obtain operating parameter information of the power battery system during operation; A data sequence construction module is used to construct a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; The event trigger detection module is used to perform event trigger analysis on the multi-dimensional operation data sequence. If it is detected that any item in the operation parameter information exceeds the corresponding preset threshold, high-frequency data sampling is started to obtain high-frequency sampling information for the corresponding period; The trend analysis module is used to analyze the change trend of high-frequency sampling information and obtain the evolution direction of the abnormal operating state of the power battery system; The encapsulation and uploading module is used to encapsulate the high-frequency sampling information and the evolution direction results, obtain the encapsulated data packet, and upload the encapsulated data packet to the remote server.

[0081] Optionally, the data sequence building blocks include: The time alignment processing submodule is used to obtain the data items containing time stamps in the operating parameter information, perform interpolation alignment processing on the data items according to a unified time step, and obtain an aligned operating parameter set; The structure filling construction submodule is used to fill the information of each dimension in the aligned operating parameter set into the specified position of the preset multidimensional data structure template based on the preset multidimensional data structure template to form a multidimensional data format; The time series integration generation submodule is used to splice and integrate multidimensional data formats in time sequence to obtain a multidimensional operation data sequence.

[0082] Optionally, the structure filling building block includes: A time index mapping unit is used to determine the corresponding time index according to the complete parameter content of each time step in the aligned operating parameter information set, and map the time index to the time axis dimension of the preset multidimensional data structure template to form a time distribution structure; The dimension filling and positioning unit is used to determine the corresponding position of the complete parameter content in the preset multidimensional data structure template according to the type and dimension of each parameter in the complete parameter content, and perform a filling operation to form a parameter distribution structure; The structure combination generation unit must have a combination rule based on the time distribution structure and the parameter mapping structure to complete the filling process of the preset multi-dimensional data structure template and form a multi-dimensional data format.

[0083] Optionally, the event trigger detection module further includes: The power supply status acquisition submodule is used to acquire the current power supply status information of the data recording path on which the high-frequency data sampling depends while executing the high-frequency data sampling; The power supply capability judgment and switching submodule is used to judge whether the current power supply capability meets the power conditions required for high-frequency data sampling based on the power supply status information. If the power supply status information does not meet the power conditions required for high-frequency data sampling, the power supply mode of the data recording path is switched to wireless power supply.

[0084] Optional, trend analysis module includes; The high-frequency sequence construction submodule is used to arrange the high-frequency sampling information in the order of acquisition time and construct a high-frequency operation information sequence corresponding to the time; The trend feature extraction submodule is used to input the high-frequency operation information sequence into the trained abnormal evolution trend recognition model to obtain trend feature information; The trend pattern recognition submodule is used to match and analyze trend feature information with a preset abnormal evolution rule library to identify trend patterns related to abnormal operating conditions of the power battery system; The evolution direction determination submodule is used to obtain the evolution direction result based on the change direction and change intensity reflected by the trend pattern, combined with the corresponding time segment and parameter indicators.

[0085] Optionally, the trend feature extraction submodule includes: A training sample construction unit is used to obtain historical high-frequency sampling information of the power battery system and abnormal evolution label information corresponding to the historical high-frequency sampling information, and to construct a supervised training sample set based on the historical high-frequency sampling information and the abnormal evolution label information; A model training execution unit is used to input the supervised training sample set into the abnormal evolution trend recognition model to be trained to obtain the model verification result; A model structure optimization unit is used to compare the model verification result with the preset model performance evaluation index. If the model verification result does not meet the preset model performance evaluation index, the structure of the abnormal evolution trend recognition model to be trained is updated, and training is continued until the model verification result meets the model performance evaluation index; The model output confirmation unit is used to stop the training of the abnormal evolution trend recognition model to be trained when the model verification result meets the preset model performance evaluation index, and obtain the trained abnormal evolution trend recognition model.

[0086] Optionally, the packaging and uploading module also includes: The data buffer management submodule is used to detect whether the current data transmission status meets the preset upload condition requirements during the process of uploading the encapsulated data packet to the remote server. If the current data transmission status does not meet the preset upload condition requirements, the encapsulated data packet will be stored in the local buffer and the corresponding time stamp and abnormal status information will be generated.

[0087] The specific definitions of a power battery data acquisition and storage device can be found in the definitions of a power battery data acquisition and storage method described above and will not be repeated here. Each module in the power battery data acquisition and storage device described above may be implemented in whole or in part through software, hardware, or a combination thereof. Each of the modules may be embedded in or independent of a processor in a computer device in hardware form, or may be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0088] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 9 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for collecting and storing power battery data is implemented.

[0089] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed: Obtain operating parameter information of the power battery system during operation; Construct a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; Perform event-triggered analysis on the multi-dimensional operating data sequence. If any item in the operating parameter information is detected to exceed the corresponding preset threshold, high-frequency data sampling is initiated to obtain high-frequency sampling information for the corresponding period. Analyze the changing trend of high-frequency sampling information to obtain the evolution direction of the abnormal operating state of the power battery system; The high-frequency sampling information and the evolution direction results are encapsulated to obtain an encapsulated data packet, which is then uploaded to a remote server.

[0090] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: Obtain operating parameter information of the power battery system during operation; Construct a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; Perform event-triggered analysis on the multi-dimensional operating data sequence. If any item in the operating parameter information is detected to exceed the corresponding preset threshold, high-frequency data sampling is initiated to obtain high-frequency sampling information for the corresponding period. Analyze the changing trend of high-frequency sampling information to obtain the evolution direction of the abnormal operating state of the power battery system; The high-frequency sampling information and the evolution direction results are encapsulated to obtain an encapsulated data packet, which is then uploaded to a remote server.

[0091] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0092] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0093] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A power battery data collection and storage method, characterized in that: The power battery data collection and storage method includes: Obtain operating parameter information of the power battery system during operation; Constructing a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; Performing event-triggered analysis on the multidimensional operation data sequence, and if it is detected that any item of the operation parameter information exceeds a corresponding preset threshold, initiating high-frequency data sampling to obtain high-frequency sampling information for the corresponding time period; Performing a change trend analysis on the high-frequency sampling information to obtain an evolution direction result of the abnormal operating state of the power battery system; The high-frequency sampling information and the evolution direction result are encapsulated to obtain an encapsulated data packet, and the encapsulated data packet is uploaded to a remote server.

2. A power battery data acquisition and storage method according to claim 1, characterized in that: The constructing of a time-synchronized multi-dimensional operation data sequence based on the operation parameter information includes: Acquire data items containing time stamps in the operating parameter information, and perform interpolation alignment processing on the data items according to a unified time step to obtain an aligned operating parameter set; Based on a preset multidimensional data structure template, filling the information of each dimension in the aligned operating parameter set into a designated position of the preset multidimensional data structure template to form a multidimensional data format; The multidimensional data formats are spliced ​​and integrated in chronological order to obtain the multidimensional operation data sequence.

3. A power battery data acquisition and storage method according to claim 2, characterized in that: The method of filling the information of each dimension in the aligned operating parameter set into a designated position of the preset multidimensional data structure template based on the preset multidimensional data structure template to form a multidimensional data format includes: Determine a corresponding time index based on the complete parameter content of each time step in the aligned operating parameter information set, and map the time index to the time axis dimension of the preset multidimensional data structure template to form a time distribution structure; Determine the corresponding position of the complete parameter content in the preset multidimensional data structure template according to the type and dimension of each parameter in the complete parameter content, and perform a filling operation to form a parameter distribution structure; According to the combination rule of the time distribution structure and the parameter mapping structure, the filling process of the preset multi-dimensional data structure template is completed to form the multi-dimensional data format.

4. The power battery data acquisition and storage method according to claim 1, characterized in that: If it is detected that any item of the operating parameter information exceeds a corresponding preset threshold, high-frequency data sampling is started to obtain high-frequency sampling information of the corresponding time period, which also includes: While executing the high-frequency data sampling, obtaining current power supply status information of the data recording path on which the high-frequency data sampling depends; Based on the power supply status information, it is determined whether the current power supply capacity meets the power condition required for the high-frequency data sampling; if the power supply status information does not meet the power condition required for the high-frequency data sampling, the power supply mode of the data recording path is switched to wireless power supply.

5. The power battery data acquisition and storage method according to claim 1, characterized in that: The performing of a change trend analysis on the high-frequency sampling information to obtain an evolution direction result of the abnormal operating state of the power battery system includes: Arrange the high-frequency sampling information in the order of acquisition time to construct a high-frequency operation information sequence corresponding to the time; Inputting the high-frequency operation information sequence into the trained abnormal evolution trend recognition model to obtain trend feature information; Matching and analyzing the trend feature information with a preset abnormal evolution rule library to identify trend patterns related to the abnormal operating state of the power battery system; The evolution direction result is obtained according to the change direction and change intensity reflected by the trend pattern, combined with the corresponding time segment and parameter indicators.

6. The power battery data acquisition and storage method according to claim 1, characterized in that: The trained abnormal evolution trend recognition model further includes: Acquire historical high-frequency sampling information of the power battery system and abnormal evolution label information corresponding to the historical high-frequency sampling information, and construct a supervised training sample set based on the historical high-frequency sampling information and the abnormal evolution label information; Inputting the supervised training sample set into the abnormal evolution trend recognition model to be trained to obtain a model verification result; Comparing the model verification result with a preset model performance evaluation index; if the model verification result does not meet the preset model performance evaluation index, updating the structure of the abnormal evolution trend recognition model to be trained, and continuing training until the model verification result meets the model performance evaluation index; When the model verification result meets the preset model performance evaluation index, the training of the abnormal evolution trend recognition model to be trained is stopped to obtain the trained abnormal evolution trend recognition model.

7. The power battery data collection and storage method according to claim 1, characterized in that: The step of encapsulating the high-frequency sampling information and the evolution direction result to obtain an encapsulated data packet, and uploading the encapsulated data packet to a remote server further includes: If, during the process of uploading the encapsulated data packet to the remote server, it is detected whether the current data transmission status meets the preset upload condition requirements, if the current data transmission status does not meet the preset upload condition requirements, the encapsulated data packet will be stored in the local buffer, and the corresponding time stamp and abnormal status information will be generated.

8. A power battery data acquisition and storage device, characterized in that: The power battery data acquisition and storage device comprises: An operating parameter acquisition module is used to obtain operating parameter information of the power battery system during operation; A data sequence construction module, configured to construct a multi-dimensional operation data sequence synchronized with time based on the operation parameter information; An event trigger detection module is used to perform event trigger analysis on the multidimensional operation data sequence, and if it is detected that any item of the operation parameter information exceeds the corresponding preset threshold, high-frequency data sampling is initiated to obtain high-frequency sampling information of the corresponding time period; a trend analysis module, configured to perform a change trend analysis on the high-frequency sampling information to obtain an evolution direction result of the abnormal operating state of the power battery system; The encapsulation and uploading module is used to encapsulate the high-frequency sampling information and the evolution direction result to obtain an encapsulated data packet, and upload the encapsulated data packet to a remote server.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the power battery data collection and storage method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the power battery data collection and storage method according to any one of claims 1 to 7 are implemented.

Images