Data analysis method and device for energy storage transmission micro-control module of new energy electric vehicle

Abstract

The invention provides a data analysis method and device for an energy storage transmission micro-control module of a new energy electric vehicle, and the method comprises the steps: carrying out the time sequence collection and preliminary filtering of multi-channel original data of a vehicle, and obtaining a key data set; executing cross-channel dynamic cache synchronization by utilizing a bandwidth prediction feedback mechanism, and carrying out time sequence cross check on adjacent data packets to generate a consistency check report; performing multi-dimensional packet tracking on the suspicious data packets in the report to form a wrong-sequence packet list; and finally, mining a trigger condition which causes the generation of a wrong-sequence packet in the multi-level correlation analysis, and outputting an abnormal early warning result. According to the invention, through effective reservation of extreme fluctuation information and accurate identification of data packet disorder and retransmission under a high-load working condition, the defect that sequence disorder caused by non-uniform bandwidth distribution and cache redundancy is difficult to find in time is overcome; and the data transmission reliability and the fault early warning capability of the energy storage system of the new energy electric vehicle can be further enhanced.

Description

Technical Field

[0001] The present invention relates to the field of automobile control technology, and in particular to a data analysis method and device for an energy storage and transmission microcontroller module of a new energy electric vehicle. Background Art

[0002] With the rapid development of the new energy vehicle industry, the integration of power battery technology and vehicle control systems continues to increase. New energy electric vehicles generate large amounts of multi-channel data during driving. To achieve more reliable energy management strategies and promptly respond to potential safety risks, vehicle manufacturers typically configure energy storage and transmission microcontroller modules between the power battery pack, temperature sensors, and the main control system to collect and transmit operating information from key vehicle components. This information includes not only the charge and discharge status of the power battery, but also data on multiple dimensions such as temperature, current, and vehicle driving conditions. Due to differences in frequency and priority among channels, existing technologies often face problems such as uneven bandwidth allocation, excessive cache load, and high-frequency sampling redundancy when performing data synchronization and analysis.

[0003] In current energy storage and transmission applications, a common practice is to perform simple aggregation and basic error detection on the data of each channel, mostly focusing on macro indicators such as packet loss rate and latency. If there is a slight timestamp offset or serial number confusion in each channel, it is usually difficult to identify it in the early stage. At the same time, when the vehicle is under high-load conditions such as acceleration and climbing, the instantaneous output power and temperature rise rate of the power battery will change significantly, and conventional filtering and analysis methods tend to ignore certain extreme values ​​or boundary anomalies with characteristic significance. Once accumulated to a certain extent, it will be very difficult to locate the fault later. In addition, there is a hidden danger of data packet disorder or retransmission under multi-channel parallel transmission. If only relying on the existing single detection mechanism, it is often difficult to detect small transmission anomalies across channels in a timely manner, and it is not conducive to accurately tracing system faults.

[0004] To further explore the energy storage status and potential failures of new energy electric vehicles in actual operating environments, the industry has attempted to improve data transmission stability by increasing bandwidth margins, expanding cache capacity, and improving sensor acquisition accuracy. However, these measures often increase hardware costs or energy consumption, and cannot fundamentally address multi-channel timing discrepancies and data congestion during vehicle operating mode transitions. Especially after long periods of operation or repeated exposure to extreme operating conditions, the transmission microcontroller module may still experience subtle missequencing or retransmissions. Summary of the Invention

[0005] The main purpose of the present invention is to solve the technical problems of uneven bandwidth allocation, excessive buffer load, and high-frequency sampling redundancy caused by existing multi-channel parallel data transmission under high-load conditions of new energy electric vehicles, which lead to extreme values ​​or boundary anomalies being easily ignored, and difficulty in timely detection of data packet disorder or retransmission; The present invention provides a data analysis method for a new energy electric vehicle energy storage and transmission microcontroller module, the data analysis method for the new energy electric vehicle energy storage and transmission microcontroller module comprising: Collecting raw data from the energy storage and transmission microcontroller module of the new energy electric vehicle, and filtering the raw data according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain a key data set; Performing cross-channel cache synchronization on the key data set, and comparing the timestamps and sequence number differences of adjacent data packets in each channel in the key data set to obtain a consistency check report; Perform multi-dimensional packet tracking on suspicious data packets marked in the consistency check report to obtain a list of out-of-order packets; According to the out-of-sequence packet list, the energy storage transmission process of the new energy electric vehicle is subjected to correlation analysis processing to obtain the triggering conditions of the out-of-sequence packet list, and the triggering conditions are used as abnormal warning analysis results of the new energy electric vehicle.

[0006] The present invention also provides a data analysis device for a new energy electric vehicle energy storage and transmission microcontroller module, the data analysis device for the new energy electric vehicle energy storage and transmission microcontroller module comprising: A data acquisition module is used to collect raw data from the energy storage and transmission microcontroller module of the new energy electric vehicle, and filter the raw data according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain a key data set; A synchronization verification module is used to synchronize the cross-channel cache of the key data set and compare the timestamps and sequence number differences of adjacent data packets in each channel of the key data set to obtain a consistency verification report; a packet tracking module, configured to perform multi-dimensional packet tracking processing on suspicious data packets marked in the consistency check report to obtain a list of out-of-order packets; The association analysis module is used to perform association analysis on the energy storage transmission process of the new energy electric vehicle according to the out-of-sequence packet list, obtain the triggering conditions of the out-of-sequence packet list, and use the triggering conditions as the abnormal warning analysis results of the new energy electric vehicle.

[0007] The data analysis method and device for the energy storage and transmission microcontroller module of the new energy electric vehicle obtain key data sets by performing time-series acquisition and preliminary filtering of the vehicle's multi-channel raw data; then, the bandwidth prediction feedback mechanism is used to perform cross-channel dynamic cache synchronization, and a time-series cross-check is performed on adjacent data packets to generate a consistency check report; then, multi-dimensional packet tracking is performed on the suspicious data packets in the report to form a list of out-of-order packets; finally, the triggering conditions that lead to the generation of out-of-order packets are mined in the multi-level correlation analysis, and an abnormal warning result is output. The present invention overcomes the defect of difficulty in timely detection of sequence disorder caused by uneven bandwidth allocation and cache redundancy by effectively retaining extreme fluctuation information under high-load conditions and accurately identifying data packet disorder and retransmission, and can further enhance the data transmission reliability and fault warning capability of the energy storage system of new energy electric vehicles.

[0008] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purposes and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0009] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a schematic diagram of a first embodiment of a data analysis method for a new energy electric vehicle energy storage and transmission microcontroller module according to an embodiment of the present invention; Figure 2 Schematic diagram of an embodiment of a data analysis device for a new energy electric vehicle energy storage and transmission microcontroller module in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0012] The terms "including," "having," and any variations thereof, as used in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to the process, method, product, or device.

[0013] To facilitate understanding of this embodiment, a data analysis method for a new energy electric vehicle energy storage and transmission microcontroller module disclosed in an embodiment of the present invention is first introduced in detail. Figure 1 As shown, this method includes the following steps: 101. Collecting raw data from the energy storage and transmission microcontroller module of the new energy electric vehicle, and filtering the raw data according to the channel characteristics, transmission priority, and vehicle operating conditions of the raw data to obtain a key data set; In one embodiment of the present invention, the raw data of the energy storage and transmission microcontroller module of the new energy electric vehicle is collected, and the raw data is filtered according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain the key data set, including: time series data collection of the power battery monitoring channel, temperature sensor channel and inertial measurement unit channel in the energy storage and transmission microcontroller module to obtain raw data with channel characteristic labels; according to the load level and transmission priority of the new energy electric vehicle under different vehicle operating conditions, the continuous sampling records in the raw data are segmented and split to obtain data to be filtered; according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data, the abnormal extreme values ​​and repeated sampling points in the data to be filtered are subjected to difference threshold judgment processing, and the abnormal extreme values ​​and repeated sampling points are deleted and merged respectively to obtain the key data set.

[0014] Specifically, the microcontroller module first assigns a unified reference time to the power battery monitoring channel, temperature sensor channel, and inertial measurement unit channel based on the vehicle's built-in high-precision clock mechanism. To ensure the acquisition of raw data tagged with channel characteristics, the system establishes a channel management table in the firmware layer to store each channel's hardware port identifier, sampling frequency, and priority information. The channel management table is loaded during power-up initialization and determines the operating mode of each channel by reading the vehicle configuration file. This table is linked to the task scheduler within the microcontroller module, which schedules acquisition tasks for each channel based on priority and the vehicle's current operating conditions. Once an acquisition task is scheduled, it triggers the hardware sampling instruction for the corresponding channel, causing the sensor or measurement module to acquire actual current, temperature, or attitude data within the specified sampling period. The data is then packaged and stored in a temporary buffer along with the current timestamp. A ring buffer structure is set up in the temporary buffer to sequentially record each piece of collected data, along with its corresponding timestamp, channel identifier, and initial validity flag. Before data packaging, the program tags each record based on the channel characteristics information stored in the channel management table. For example, when the power battery monitoring channel collects information such as voltage and current, it is labeled with the "Battery" channel tag and the system clock value at the time the record was generated. The temperature sensor channel is assigned the "Temperature" tag, and the inertial measurement unit channel is assigned the "IMU" tag. Channel attribute tags are not limited to names but also include fields such as priority and hierarchical level, allowing for more complete records of high-priority data during subsequent filtering. Certain channels also add custom parameters based on vehicle commissioning configurations, such as recording the corresponding battery pack serial number or specific number of cells in the battery channel. During data output, each record's timestamp and channel tag are stored in a unified structure and written to a reusable queue data structure. This queue data structure is associated with a temporary buffer at vehicle startup. When the queue is full, an overflow handler is invoked to prevent the loss of critical data. The resulting multi-channel raw data includes high-precision timestamps, channel tags, and basic measurement values. Subsequent modules can directly determine the record type and priority based on the tags, thereby implementing appropriate processing strategies for different scenarios.

[0015] Specifically, after receiving multi-channel raw data, the system begins segmenting the continuously sampled records based on the load level of the new energy electric vehicle under different operating conditions and the transmission priority of each channel. This requires maintaining a mapping between operating conditions and loads in the upper-level logic of the microcontroller module to determine whether the vehicle is accelerating, braking, or driving steadily. This mapping is stored in a global variable generated during the interaction between the vehicle control unit and the power battery management system and is periodically read after the microcontroller module starts. When the vehicle acceleration exceeds a certain threshold or the battery output power increases significantly, the system identifies this period as a high-load condition. When the vehicle is detected as cruising or idling steadily, it is identified as a low-load condition. To segment, the program chronologically browses the multi-channel raw data queue and extracts the timestamp, channel label, and priority value of each record. If multiple consecutive records are found to fall within the same load range and have similar channel priorities, the system groups these records into a single segment and records the segment's start and end timestamps. During segmenting, the priority value is used to determine the urgency of each record. For example, if the battery discharge current is a high-priority channel during high-load conditions, data from a channel with a significantly different priority during that period is placed in a separate segment for management. This allows the high-priority channel data to be processed independently during subsequent filtering, preventing it from being overwhelmed by the massive amount of information from the temperature or inertial measurement unit channels. In high-load scenarios such as acceleration or hill driving, the system performs more granular segmentation of records. Whenever channel data fluctuates significantly and remains above a given threshold for a short period of time, the program further divides that period into sub-intervals, accurately assigning each record to the appropriate segment. During periods of stable driving or vehicle idling, the segmentation strategy is more relaxed, merging adjacent records within a longer timeframe. Once this is complete, the program labels each segment "data to be filtered," indicating that it will be subject to difference threshold determination and exception handling in the next stage. Each segment also contains additional information, such as the system clock values ​​for the start and end of the segment, a high-load or low-load status indicator, and the number of internal record entries. Ultimately, all multi-channel raw records are split into segments based on load level and priority, and stored in a segment management table. This segment management table can be associated with specific data content in the query queue and provide a query interface for the next stage of threshold determination.

[0016] Specifically, after segmentation is complete, the system reads the "data to be filtered" segment by segment according to the segment management table and applies a difference threshold test to eliminate abnormal extreme values ​​and duplicate sampling points, ultimately outputting the final key data set. During this process, the program first applies threshold rules based on each segment's high-load or low-load label. For high-load segments, thresholds are often more relaxed because the battery current and temperature can fluctuate significantly during rapid acceleration or on a slope. Overly strict thresholds could result in excessive deletion of critical information. For low-load segments, stricter anomaly detection rules are employed, allowing for rapid elimination of measurements that significantly deviate from the normal range to prevent excessive data interference and impact analysis. During the specific test, the program accesses each record within the segment and compares its measured value against a predetermined upper or lower limit. If a temperature channel value is observed that is significantly outside the acceptable range, the system considers the record to be extreme noise and deletes it. If the battery current sample has multiple, similar values ​​within a very short interval, and the difference is less than the set threshold, the program merges the sample, retaining only the first sample and noting the size of the merged segment. Sudden high current fluctuations consistent with vehicle acceleration or braking, if they exceed the average level by a certain multiple, may also be recorded separately to facilitate subsequent identification of potential misordering or retransmissions. For inertial measurement unit channel data, the program compares the attitude change rates of consecutive samples and, combined with the load identifiers used during segmentation, determines whether any jitter is invalid. If deemed to be unreasonable jitter, the program deletes the record. During this process, the system assigns a new filter tag to each filtered or merged data entry, recording the reason for deletion or retention and whether it was merged into a merged entry due to repeated sampling. These tags, along with the filtering results, are written to an intermediate cache file to facilitate subsequent verification of the data processing process. After completion, the retained entries are output as a key data set, retaining their original channel tags and timestamp fields. This ensures accurate location within the overall data stream during transmission synchronization or misordering detection. The key data set thus contains only valid records that are valuable for subsequent analysis, significantly reducing the proportion of invalid or redundant data.

[0017] 102. Perform cross-channel cache synchronization on the key data set, and compare the timestamps and sequence number differences of adjacent data packets in each channel in the key data set to obtain a consistency check report; In one embodiment of the present invention, the cross-channel cache synchronization of the key data set and comparing the timestamp and sequence number differences of adjacent data packets in each channel in the key data set to obtain a consistency verification report include: performing bandwidth prediction based on the historical bandwidth occupancy of each channel in the key data set to obtain a bandwidth demand prediction result in a future time window; based on the bandwidth demand prediction result, dynamically allocating priorities for data packets in each channel in the key data set, and adjusting the cache space of each channel according to the allocation result to obtain a cross-channel cache allocation scheme; according to the cross-channel cache allocation scheme, performing cache synchronization processing on the key data set to obtain a synchronized multi-channel data stream; performing difference calculation on the timestamps and sequence numbers of adjacent data packets in the multi-channel data stream to obtain a timestamp difference sequence and a sequence number difference sequence; according to a preset timestamp tolerance threshold and sequence number tolerance threshold, performing anomaly detection on the timestamp difference sequence and the sequence number difference sequence, marking suspicious data packets, and obtaining a consistency verification report.

[0018] Specifically, the system uses a preset bandwidth prediction feedback mechanism to perform bandwidth prediction calculations based on the historical bandwidth usage of each channel in the key data set. First, a bandwidth usage statistics table is maintained in the memory of the energy storage transmission microcontroller module. Each channel periodically writes historical records to this table, recording items such as the channel's peak and average traffic flow during various time periods. The program then uses methods such as weighted sliding average or exponential smoothing to numerically fit the statistics table, taking into account the traffic fluctuation characteristics of vehicles in high-load scenarios (such as acceleration and hill climbing). This method generates a bandwidth demand curve for each channel in the future time window and estimates the occupancy probability for channels of different priority levels. The bandwidth prediction feedback mechanism periodically re-retrieves the actual traffic data for each channel and adjusts the prediction results based on the latest occupancy data, ensuring that the prediction curve more closely reflects current load fluctuations. After calculations, the program generates a bandwidth demand forecast report in memory, listing the future bandwidth demand levels for each channel, to facilitate dynamic priority allocation in subsequent steps.

[0019] Specifically, based on the bandwidth demand forecast results, the system dynamically prioritizes packets across channels in the critical data set and adjusts the cache space allocated to each channel based on the allocation results, thereby forming a cross-channel cache allocation plan. The program then accesses the peak channel occupancy levels listed in the bandwidth demand forecast report and, combined with weighting factors for different operating conditions such as vehicle acceleration and braking, compares the urgency and historical reliability of each channel. For channels with high occupancy levels and high priority within the forecast window, the program allocates larger cache capacity. For channels with average occupancy levels, the cache queue is appropriately reduced without compromising data integrity. The system also records a cache mapping structure in the allocation results, establishing a correspondence between the cache queue upper limit and remaining space for each channel. If bandwidth is subsequently exceeded, the program dynamically adjusts the queue capacity based on this mapping structure. After completing this allocation, the system outputs a cross-channel cache allocation plan, specifying the available cache range for each channel in the next phase.

[0020] Specifically, according to the cross-channel cache allocation scheme, cache synchronization is performed on critical data sets, resulting in synchronized multi-channel data streams. The program compares the priority and timestamp of each packet with the allocation scheme, ensuring that packets from high-priority channels receive immediate write slots in the corresponding cache queues and maintaining consistent arrival order within the storage structure. If the allocated cache space for a channel is small, older, lower-priority packets from the occupied queue are retrieved when a packet is written and moved to a secondary storage area or replaced to prevent critical data from being squeezed out of the cache. After queue adjustments are complete, the system scans records from different time periods one by one and fuses multi-channel data from the same time period into a single, continuous data stream based on a unified time sequence. The multi-channel data streams now have consistent start and end timestamps in memory, enabling rapid location of corresponding time sequences during the next stage of difference calculation. This fusion process is performed frequently under high-concurrency conditions to maintain synchronization of multi-channel data.

[0021] Specifically, the timestamps and sequence numbers of adjacent data packets in the synchronized multi-channel data stream are calculated to obtain a timestamp difference sequence and a sequence number difference sequence. The system traverses the records of each channel in the data stream and performs a subtraction operation on the adjacent records of each channel. The timestamp of the latter record is subtracted from the timestamp of the previous record and stored in the timestamp difference sequence. The sequence number of the latter record is subtracted from the sequence number of the previous record and stored in the sequence number difference sequence. If there are multiple consecutive records in the same time period for a channel, the program calculates the difference between each pair and arranges the results in order. To avoid subtle errors caused by high-frequency sampling, the program uses high-precision integers or fixed-point arithmetic to record time differences and sequence differences. After this calculation is completed, a set of timestamp difference sequences and sequence number difference sequences will be generated, providing a quantitative basis for subsequent anomaly detection.

[0022] Specifically, according to the preset timestamp tolerance threshold and sequence number tolerance threshold, the timestamp difference sequence and sequence number difference sequence are detected for anomalies one by one, suspicious data packets are marked, and a consistency verification report is obtained. The program reads the timestamp and sequence number thresholds of each channel from the configuration file and compares each difference element. If the timestamp difference exceeds the given upper limit, the data packet containing the record will be marked as a delay anomaly; if the sequence number difference is negative or the jump value is far beyond the reasonable range, an out-of-order anomaly prompt will be written in the tag field. For batch anomalies discovered during multi-channel synchronization, the program will add cross-channel anomaly tags to identify certain global synchronization problems. The marking results include the anomaly type, occurrence time, and channel information, and are output to the data management module in the form of a consistency verification report.

[0023] 103. Perform multi-dimensional packet tracking on suspicious data packets marked in the consistency check report to obtain a list of out-of-sequence packets; In one embodiment of the present invention, the multi-dimensional packet tracking processing of the suspicious data packets marked in the consistency check report to obtain a list of out-of-sequence packets includes: extracting relevant data packets within a preset time window before and after the channel to which the suspicious data packet belongs according to the marking information in the consistency check report to form a local data packet sequence; calculating the time interval and serial number interval between the suspicious data packet and the adjacent data packets according to the local data packet sequence to obtain an interval feature vector; calling a preset multi-dimensional judgment algorithm to classify the interval feature vector, determine the out-of-sequence type of the suspicious data packet, and obtain an out-of-sequence judgment result; and performing context correlation analysis on the out-of-sequence judgment result in combination with the vehicle operating conditions and channel load status to obtain a list of out-of-sequence packets.

[0024] Specifically, the program first reads the consistency check report output by the previous step and scans each packet record marked as exceeding the timestamp or sequence number tolerance. Each record contains the channel tag carried by the packet at the time of acquisition, the detailed type of anomaly, and the corresponding timestamp and sequence number. The program then creates a temporary index structure in memory to summarize the key elements of these suspicious packets. The channel tag distinguishes the source, such as the power battery monitoring channel, temperature sensor channel, or inertial measurement unit channel. The timestamp is used to locate the packet's position in the global data flow during subsequent packet tracing. The sequence number is used to identify any jumps or retransmissions. To facilitate subsequent matching, the index structure assigns a unique identifier to each suspicious packet and retains its reference line number from the consistency check report, making it easier to locate the original record. This index table is implemented using data structures such as hashes or balanced trees to ensure efficient matching and access when querying or adding index items. Once the system has sequentially written all anomaly entries to the index table, the entire process completes the extraction and preliminary organization of suspicious packets. The rationale behind creating an index table is that out-of-order or retransmission issues often manifest only at specific times and channel combinations. Therefore, this method is necessary to centralize scattered suspicious records, enabling quick retrieval of contextual data associated with these records in subsequent stages. The program also notes the time range of the suspicious packets in an additional field of the index table, useful for identifying high-load conditions or near-channel switching moments. This way, the index table not only contains core information about suspicious packets but also lays the necessary logical connections for constructing local packet sequences in the next stage. The entire indexing process operates on a one-way read of the consistency check report and does not modify or affect the structure of the report itself. It simply reorganizes the flagged entries in the report into a collection of suspicious packets that is easier to process. Once all suspicious packets are registered in the index table, they can be tracked in depth, enabling more detailed analysis of each record's channel and time period.

[0025] Specifically, the program sequentially reads the timestamp and channel tag of each suspicious packet from an index table, then searches for records adjacent to that timestamp in the multi-channel synchronized data stream or raw data cache. During the search, the system locates the time series data for the corresponding channel based on the channel tag and groups records whose timestamps fall within a certain range before or after the suspicious packet into a temporary sequence structure. The size of this range is determined by a preset time window, the length of which is typically determined by the vehicle's maximum rate of change within a single sampling cycle or the high-frequency fluctuation characteristics of the channel. For example, the power battery monitoring channel has a faster sampling rate during rapid acceleration, so a relatively tight window can be set, while the inertial measurement unit channel may require a wider period to capture a coherent motion trajectory during sudden posture changes. The program then concatenates the records whose timestamps fall within the same window as the suspicious packet in chronological order, retaining the original sequence number information for each record. This resulting localized packet sequence reflects the context of the suspicious packet at several sampling moments before and after it, including the time difference between adjacent packets, sequence number increments, and background indicators such as the vehicle's operating mode switching and channel load during this interval. The reason for extracting these local sequences is that out-of-order or retransmissions often occur during brief periods of high concurrency or bandwidth constraints. Simply examining a suspicious packet isn't enough to confirm a true anomaly; only by combining several preceding and subsequent records can we infer the normal evolution of sequence numbers within the channel. After extraction, the program creates a local sequence storage unit for each index table entry, recording the time and sequence differences between the suspicious packet and its adjacent records.

[0026] Specifically, the program sequentially scans the local sequence corresponding to each suspicious packet and calculates the timestamp and sequence number differences between adjacent records in chronological order. For each local sequence, the program treats the suspicious packet as the central node and searches forward and backward for several records, performing a subtraction operation on each record: subtracting the timestamp of the previous record from the timestamp of the suspicious packet, or subtracting the timestamp of the next record from the timestamp of the suspicious packet. The sequence number difference of the suspicious packet is calculated using the same logic. The generated time intervals and sequence number intervals are stored in fixed-point or high-precision floating-point format and arranged sequentially within each local sequence to form a set of feature vectors. This is done because packet delivery speeds vary across channels and under different operating conditions. Relying solely on the suspicious packet's data items makes it difficult to determine the type of anomaly. Time intervals and sequence number intervals can more intuitively indicate whether there are skips, retransmissions, or unusual transmission orders. The program also records the suspicious packet's relative position to the entire sequence, such as whether it is at the beginning, middle, or end of the local sequence, as well as the priority assigned to the records on either side. After generating the feature vector, the program temporarily stores it in the extended field of the suspicious data packet with an identifier indicating which local sequence the feature vector belongs to.

[0027] Specifically, a pre-defined multi-dimensional decision algorithm is invoked to classify the interval feature vectors, determine the type of missequence for the suspicious packet, and generate a missequence determination result. The program loads a set of decision models based on statistics or machine learning and inputs the feature vector for each suspicious packet into the corresponding model interface. The decision algorithm considers whether the time intervals within the local sequence of suspicious packets are normally or exponentially distributed, whether the sequence number intervals show large jumps or negative increments, and whether there are inconsistencies in the priority or load relationships between channels. If both the time interval and the sequence number interval are detected to be outside the normal range, and the channel containing the packet is highly loaded during that time period, the algorithm outputs a label labeled "highly likely missequence" or "duplicate transmission." If only the time interval anomaly is detected but the sequence number difference is small, the algorithm labels it "minor missequence." If the sequence number jump is large but the timestamp deviation is not significant, the anomaly is classified as "retransmission due to channel congestion." The model also considers the channel's operating mode and priority allocation strategy during the classification process to determine whether these interval characteristics are consistent with the vehicle's main control system's transmission pattern during that time period, thereby generating a more accurate missequence determination result. After completing the judgment, the system will write these preliminary results back to the extended field of the suspicious data packet index table, along with parameters such as classification probability or credibility to indicate the severity of the misordering or duplication phenomenon.

[0028] Specifically, the system performs contextual correlation analysis on the out-of-sequence determination results, combining vehicle operating conditions and channel load status. To generate the out-of-sequence packet list, the system retrieves real-time or recently recorded operating condition data from the vehicle control unit, including fields such as acceleration, speed, battery temperature, or current, and feeds these fields, along with channel load status, into the correlation analysis module. The system then examines each suspicious packet's anomaly label from the initial assessment and compares the vehicle load level during the time period with the algorithm's out-of-sequence conclusion. If the timestamp and sequence number jumps during extremely high channel load conditions are consistent with high-priority congestion, the system assigns a "high-load out-of-sequence" label. If the vehicle is operating steadily but the channel spacing is unusually high, this is considered a more serious transmission anomaly. Furthermore, the system further verifies records identified as "duplicate transmissions" or "minor out-of-sequences." If the vehicle's operating conditions during this period were relatively stable, improper cache allocation may be a possible cause. The rationale for contextual correlation analysis is that different vehicle operating conditions can produce distinct communication characteristics. High acceleration or extreme temperature environments can induce severe congestion or packet loss, while anomalies in stable conditions may indicate other design deficiencies in the microcontroller or channel configuration. After completing the correlation analysis, the system updates the final misordering type in the corresponding entry in the suspicious packet index table and compiles all packets determined to be misordered or duplicated into a misordered packet list. This list includes the suspicious packet's label, the vehicle operating conditions during the time period, the channel load level, and the misordering type determined by the algorithm.

[0029] Furthermore, the combination of the vehicle operating condition and the channel load status to perform context correlation analysis on the out-of-sequence determination result to obtain the out-of-sequence packet list includes: obtaining the channel load status of each channel in the corresponding time period, and performing time series segmentation processing on the vehicle operating condition to identify the key time points of the operating condition change to obtain an operating condition change sequence; according to the operating condition change sequence, performing operating condition matching on each suspicious data packet in the out-of-sequence determination result to obtain an out-of-sequence candidate list with an operating condition mark; according to the channel load status, calculating the probability score of out-of-sequence occurrence for each data packet in the out-of-sequence candidate list to obtain an out-of-sequence probability vector; according to a preset out-of-sequence determination threshold, screening the out-of-sequence probability vector, marking the data packets above the threshold as confirmed out-of-sequence, to obtain the out-of-sequence packet list.

[0030] Specifically, the channel load status for each channel during a corresponding time period is obtained. The system first collects real-time data from the power battery output current, temperature sensor values, and the acceleration or angular velocity of the inertial measurement unit (IMU) during vehicle operation. Based on the vehicle's speed and main control commands, it identifies whether the vehicle is currently accelerating, braking, or in steady driving. To determine the channel load status, the program performs statistical operations on key load parameters for each channel at fixed time intervals. For example, it records the maximum and average currents during the interval, or calculates the temperature rise rate for the temperature channel and the frequency of attitude changes output by the IMU. This data is combined with the timestamps of suspicious packets identified in the previous stage to determine the average bandwidth utilization and transmission frequency of each channel at the time the suspicious packets occurred. During this process, the system invokes the buffer management module within the firmware to retrieve load monitoring records adjacent to the timestamp of the suspicious packets. Aggregate analysis is then performed across multiple consecutive time periods, forming a matrix structure with rows and columns corresponding to time periods and channels, respectively. The row coordinates of the load status matrix can be arranged in chronological order, while the column coordinates can distinguish different channel categories, such as power battery, motor control, and temperature or IMU. Each matrix element thus contains the load level of a channel during a specific time period. If it is above a preset threshold, it is displayed as high load; otherwise, it is recorded as low or medium load. The reason for this is that multi-channel data exhibits a multi-dimensional synchronous evolution relationship during vehicle operation. Only by clearly listing the load conditions of each channel can we combine the transmission anomalies of suspicious data packets under different load conditions in subsequent steps to make inferences. After the matrix is ​​constructed, the program will store it in a two-dimensional array or sparse matrix structure in memory, with the start and end timestamps of each time period, providing basic data for subsequent operating condition analysis and missequence probability calculation.

[0031] Specifically, the system segments the vehicle's operating conditions into time series, identifying key time points where the conditions change, and generating a condition change sequence. After the vehicle starts, the system continuously monitors data on multiple dimensions, including vehicle speed, acceleration, and battery output power. It periodically updates a data structure called a "condition sequence cache." This cache records the vehicle's operating condition at each moment, such as acceleration, deceleration, steady driving, or steep incline climbing. When the system detects a significant sudden change in vehicle speed or acceleration, or a brief shift in battery current from a low to a high load range, it marks the current moment in the cache as a condition change point. To centrally present these change points, the program periodically iterates through the condition sequence cache, assigning the time range between two consecutive change points to a segment. Each segment is accompanied by the primary operating condition type and average parameter value for that period. Once all segments are complete, the system merges these segments to generate a chronological condition change sequence, with each element containing both the time the change occurred and the condition label for that moment. The reason for this is that different vehicle operating conditions often correspond to different communication traffic patterns. If a vehicle experiences out-of-order or retransmissions during high-load periods, it is most likely related to frequent data transmission or buffer congestion. However, out-of-order transmissions during low-load periods indicate anomalies in channel allocation or scheduling. The operating condition change sequence is timestamped and associated with suspicious packets, allowing the precise location of the operating condition segment in the subsequent matching phase. This sequence is output with an index list, indicating the start and end times of each operating condition switch and the corresponding status flag.

[0032] Specifically, according to the operating condition change sequence, each suspicious data packet in the misorder determination result is matched with the operating condition to obtain a misorder candidate list with operating condition tags. In order to achieve this process, the system first reads all switching moments from the operating condition change sequence of the previous stage and stores them in a one-dimensional index in chronological order. Subsequently, for each suspicious data packet in the misorder determination result, the program extracts its timestamp information and uses binary search or other fast retrieval methods to find out in which operating condition segment the timestamp is located in the operating condition change sequence. If it is found that the timestamp of the suspicious data packet falls exactly in the interval between two operating condition switches, the data packet is assigned a working condition label corresponding to the interval, such as "high-speed driving" or "hill load". If the timestamp of the suspicious data packet coincides with the operating condition switching point, the system will regard it as a special case at the junction and retain an identifier indicating that it has the potential impact of dual operating conditions. After completing the working condition matching, the program will have a new data structure, namely a list of misordered candidates with working condition tags. Each record in the list will add a "current working condition" field based on the misorder determination result to indicate the driving state or load scenario in which the suspicious data packet was identified as suspicious. The reason for doing this is that by matching suspicious data packets with vehicle working conditions, it is possible to better distinguish the types of transmission anomalies under high load and low load conditions, and lay a data foundation for further quantitative analysis based on the channel load status. After the new misordered candidate list is generated in the memory, it will be sorted according to the order in which the data packets appear and retained in the shared memory module.

[0033] Specifically, for each packet in the out-of-order candidate list with a working condition tag, a probability score for a misorder is calculated based on the channel load status, resulting in a misorder probability vector. The system first loads the channel load status into memory and reads the load value between the channel and the time period of the suspicious packet. It then accesses the channel and time information of each suspicious packet in the out-of-order candidate list, matches the timestamp with the channel tag, and identifies the corresponding cell in the load status matrix. The load value can be the average bandwidth usage within a period or a statistically determined peak traffic value. The program normalizes these load values ​​and inputs them into the probability calculation module, along with the high or low load category specified by the working condition tag. The probability calculation module then combines the load differences between adjacent packets, the priority configuration, and the previous misorder type determination results to derive a numerical value representing the probability of the suspicious packet being misordered under these channel load conditions. This value is also correlated with the vehicle's acceleration or current peak during that period. For example, in a high-load scenario, the closer the load is to the limit, the greater the misorder probability, and vice versa. Finally, the program outputs a probability score for each suspicious packet, and these scores are arranged to form a misorder probability vector. After generating this vector, the system will add a column to the out-of-order candidate list to record the corresponding probability score and the reference load status to indicate the weight level of the suspicious packet out-of-order.

[0034] Specifically, the out-of-order probability vector is filtered based on a preset out-of-order threshold. Packets above the threshold are marked as confirmed out-of-order, resulting in a list of out-of-order packets. The system pre-defines several threshold ranges in the configuration file to distinguish between low-probability and high-probability out-of-order conditions, as well as critical thresholds. After obtaining the out-of-order probability vector, the program compares the probability score of each suspicious packet against the threshold. If the score exceeds the highest threshold, the packet is explicitly marked as "confirmed out-of-order." If the score falls below the lowest threshold, the packet is excluded from the out-of-order condition. Any packets in between are recorded with a reserved flag for subsequent manual or log review. After this screening step, the program stores all "confirmed out-of-order" entries in the out-of-order packet list and adds a description of the load status and operating condition tags. This allows for rapid identification of the most severe out-of-order points during subsequent analysis or alarming. This list is often retained within the system for a relatively long period of time to assist engineers or control strategy modules with fault diagnosis or scheduling optimization.

[0035] 104. According to the out-of-sequence packet list, perform correlation analysis on the energy storage transmission process of the new energy electric vehicle to obtain the triggering conditions of the out-of-sequence packet list, and use the triggering conditions as the abnormal warning analysis results of the new energy electric vehicle.

[0036] In one embodiment of the present invention, the energy storage and transmission process of the new energy electric vehicle is subjected to association analysis processing based on the out-of-order packet list to obtain the triggering conditions of the out-of-order packet list, and the triggering conditions are used as the abnormal warning analysis results of the new energy electric vehicle, including: performing time clustering analysis on the out-of-order packets in the out-of-order packet list to obtain out-of-order event clusters; extracting vehicle operating parameters within a corresponding time period based on the out-of-order event clusters, and performing time series analysis on the vehicle operating parameters to obtain a state transition sequence related to the out-of-order event; calling a preset association rule mining algorithm to extract association rules between out-of-order events and system state changes based on the state transition sequence to obtain the triggering conditions of the out-of-order packet list, and using the triggering conditions as the abnormal warning analysis results of the new energy electric vehicle.

[0037] Specifically, a temporal clustering analysis is performed on the out-of-order packets in the out-of-order packet list to identify the temporal distribution characteristics of out-of-order events and form clusters of out-of-order events. The system first retrieves the timestamp information carried by each out-of-order packet in the out-of-order packet list and sorts all out-of-order packets chronologically. The program then quantifies the differences between these timestamps using a sliding window or density-based clustering algorithm, identifying clusters of out-of-order events occurring within similar time periods. When the algorithm detects a high frequency of out-of-order events at certain consecutive time points or time intervals, forming a cluster on the time axis, these out-of-order packets are grouped together in the same temporal cluster. This clustering process can identify periods of concentrated out-of-order events within long-term data records, providing an objective basis for subsequent correlation analysis. When the program scans the timestamps of each out-of-order packet, it excludes sporadic out-of-order events that occur at isolated moments and have no clear correlation with the overall temporal distribution. If the number and duration of these scattered events meet certain algorithmic thresholds, they are retained as separate, small temporal clusters. After clustering is completed, the system outputs the data structure of the out-of-order event cluster, where each time cluster corresponds to a time range and contains a list of corresponding out-of-order packets, so that the vehicle operating parameters and status information associated with each time cluster can be quickly queried in the next stage.

[0038] Specifically, the system extracts vehicle operating parameters within the corresponding time period based on the out-of-sequence event clusters, constructs a multidimensional state vector, and ultimately generates the state transition sequence associated with the out-of-sequence event through time series analysis. After obtaining the out-of-sequence event clusters, the program extracts the start and end timestamps of each time cluster and retrieves key measurements within that time range from the vehicle's overall data record, such as power battery current, temperature sensor data, and vehicle speed and rotational speed. The program then performs a segmented aggregation of these parameters, organizing the data within the same time cluster into several slices, each corresponding to a finer time interval. This allows for multi-dimensional observation of the linkage between current, temperature, and vehicle operating conditions. To facilitate further analysis, the system organizes the parameters within each slice into a multidimensional vector and annotates the time point at which the out-of-sequence event occurred. This allows for time series comparison of the parameter values ​​and change directions at several sampling points before and after the out-of-sequence event. The program then uses methods such as sliding windows or differential statistics to screen for information such as the slope and peak value of the parameter changes near the out-of-sequence point and records any state transitions that are highly correlated with the out-of-sequence event. If certain parameters have begun to rise or fall rapidly before the occurrence of a missequence, it is reasonable to regard them as early signals that may induce a missequence event.

[0039] Specifically, a preset association rule mining algorithm is invoked to extract association rules between out-of-sequence events and system state changes based on the state transition sequence, thereby generating trigger conditions for the out-of-sequence packet list. These trigger conditions are then used as the abnormal warning analysis results for new energy electric vehicles. After receiving the state transition sequence output from the previous stage, the system inputs the multidimensional vector sequence within each time cluster into the association rule mining engine. The engine scans the state transition sequence for changes in fields such as temperature, current, and vehicle speed, as well as the time points and intensity markers of out-of-sequence events. Using metrics such as support and confidence in the mining algorithm, the engine assesses the degree of correlation between various parameter combinations and out-of-sequence events. If, under certain conditions, a specific set of parameter change patterns and out-of-sequence events are observed to occur multiple times simultaneously, the algorithm outputs a rule indicating that the probability of an out-of-sequence event is high under this parameter combination. After the rule is generated, the program selects the most representative trigger conditions based on preset minimum support or minimum confidence thresholds. The trigger conditions include the range of each parameter, the vehicle's operating state, and the distribution of out-of-sequence events within the time cluster, forming a conclusion that can be easily referenced by operation and maintenance and safety monitoring units. The system writes these trigger conditions into a unified abnormal warning analysis result and compares it with the vehicle's current operating mode or the state it may enter in the future. If a similar environment or parameter combination is detected again, it can be reported immediately or prompted with the risk of missequence.

[0040] Furthermore, the method of extracting vehicle operating parameters within a corresponding time period based on the out-of-sequence event clusters, and performing time series analysis on the vehicle operating parameters to obtain a state transition sequence related to the out-of-sequence events includes: extracting vehicle operating parameters of a preset time length before and after the corresponding time period from a vehicle state monitoring system based on the out-of-sequence event clusters; segmenting the vehicle operating parameters using a sliding time window method, extracting statistical features and frequency domain features within each time window, and obtaining a multidimensional state vector; applying a principal component analysis algorithm to the multidimensional state vector to obtain a reduced-dimensional feature vector, and performing time series alignment on the reduced-dimensional feature vector using a dynamic time warping algorithm to obtain a key parameter change sequence; and mapping the key parameter change sequence into discrete state identifiers according to a preset state transition rule library, and connecting them in chronological order to obtain a state transition sequence related to the out-of-sequence events.

[0041] Specifically, based on the out-of-sequence event clusters, the system extracts vehicle operating parameters for a preset time period before and after the corresponding time period from the vehicle condition monitoring system to obtain a raw parameter time series. This series is then preprocessed to obtain a standardized parameter series. After identifying the out-of-sequence event clusters, the system retrieves multi-dimensional data from the vehicle condition monitoring system, including power battery current, temperature sensing values, and vehicle speed and acceleration, based on the start and end timestamps of each time cluster. During the search, a preset time period is reserved on both sides of the time cluster to capture state changes before and after the event. The program then performs denoising and complementation on the retrieved raw parameter time series and converts them to the same time base to avoid sampling misalignment between high- and low-frequency sensors. The program then performs normalization on each parameter channel, converting values ​​of different dimensions or ranges to the same numerical range. Normalization methods such as Z-score or Min-Max are typically used to map each parameter to a more comparable range. This allows the numerical values ​​of multiple channels at the same time to directly reflect their relative weight in the overall state. Data preprocessing also includes removing possible extreme pulse anomalies and performing interpolation or smoothing of nearby values ​​for missing records. This step results in a standardized parameter sequence that no longer contains peak amplitudes that differ significantly across channels, allowing subsequent sliding window and feature extraction to proceed along the same dimension.

[0042] Specifically, a sliding time window method is used to segment the vehicle operating parameters. Within each time window, statistical and frequency domain features are extracted to generate a multidimensional state vector. The system sets a window length based on the overall sampling period of the vehicle operating data and the likely duration of the out-of-sequence event. The window then slides across the standardized parameter sequence from left to right with a fixed step size. Each slide captures a time slice, and the temperature, current, speed, and other values ​​within that time slice are treated as an independent segment. To map this segment into a processable multidimensional vector, the system calculates statistical features within the segment, such as mean, variance, maximum, and minimum values. It also performs a short-time Fourier transform or wavelet transform to obtain frequency domain features, such as energy distribution and dominant frequency components. When a segment contains multiple channels, these features from all channels are concatenated into a higher-dimensional vector and stored in a temporary feature buffer. The step size affects the amount of overlap between windows. A smaller step size results in a higher degree of overlap between adjacent windows, facilitating the capture of short-term state transitions. A larger step size reduces computational complexity but also reduces the accuracy of state transitions. After processing each window, the system shifts right by one step until the end of the normalized parameter sequence is completely covered. This ultimately results in a multidimensional sequence of state vectors. These vectors contain both statistical indicators in the time domain and energy or harmonic characteristics in the frequency domain, which can be used in subsequent dimensionality reduction and time alignment steps.

[0043] Specifically, the system applies principal component analysis (PCA) to the multidimensional state vector to obtain reduced eigenvectors. These reduced eigenvectors are then time-aligned using a dynamic time warping (DTW) algorithm, ultimately yielding a sequence of key parameter changes. After collecting the multidimensional state vectors, the system first performs PCA on the vector set to reduce algorithmic overhead and noise interference. By calculating the covariance matrix and extracting eigenvectors, the system projects multiple channels and features onto a small number of principal components. This reduces the originally high-dimensional vectors to a relatively low dimensionality while preserving key variance information, thereby reducing computational complexity. After dimensionality reduction, the system organizes the new vector set into chronological order and applies a dynamic time warping (DTW) algorithm for time alignment. Time alignment is necessary because when a vehicle is operating under high load or non-stationary conditions, the evolution rate of parameters varies, making it easy to misalign features when comparing them with the same step size. DTW adaptively aligns time axes to better align similar trends, thereby accurately matching key turning points between different windows. After alignment, the feature vectors are arranged in time sequence to form a key parameter change sequence. This sequence can intuitively reflect the evolution of the main components from window to window. If a large mutation occurs in a certain period of time, a clear jump can be detected in the dimensionality reduction space.

[0044] Specifically, based on a preset state transition rule library, the key parameter change sequence is mapped into discrete state identifiers and connected in chronological order to generate the state transition sequence associated with the out-of-sequence event. After loading the key parameter change sequence, the system matches each reduced feature vector against the rule library defined in the configuration file to determine the corresponding operating state type in the feature space. The rule library contains several thresholds and logical conditions to distinguish different operating modes. For example, a high temperature and medium current range is defined as a medium load state, while a high current and a significant increase in high-frequency energy distribution is considered a high load state. As the program traverses the key parameter change sequence, if it detects that a feature vector meets the conditions of a specific rule, it adds a discrete state identifier to the sequence for that period, such as "normal fluctuation" or "peak fluctuation." As the time sequence progresses, different vectors fall into different state intervals. The program then connects these discrete states in sequence to generate a sequence of markers reflecting the state transitions before and after the out-of-sequence event. The final output of the state transition sequence associated with the out-of-sequence event consists of several "period-state" segments, each with a relevant time window and state label. These segments clearly depict the evolution of the vehicle's principal components before and after the out-of-sequence event. At this stage, the system also adds additional features to the discrete state identifiers, such as stability or deviation, to distinguish the severity of normal and abnormal states.

[0045] In this embodiment, the key data sets are obtained by performing time-series acquisition and preliminary filtering of the vehicle's multi-channel raw data. Subsequently, a bandwidth prediction feedback mechanism is used to perform cross-channel dynamic cache synchronization, and a time-series cross-check is performed on adjacent data packets to generate a consistency check report. Multi-dimensional packet tracking is then performed on the suspicious data packets in the report to form a list of out-of-order packets. Finally, the triggering conditions that lead to the generation of out-of-order packets are mined in a multi-level correlation analysis, and an abnormal warning result is output. The present invention overcomes the problem of difficulty in timely detection of sequence disorders caused by uneven bandwidth allocation and cache redundancy by effectively retaining extreme fluctuation information under high-load conditions and accurately identifying data packet disorder and retransmission. This can further enhance the data transmission reliability and fault warning capabilities of the new energy electric vehicle energy storage system.

[0046] The above describes the data analysis method of the energy storage and transmission micro-control module of the new energy electric vehicle in the embodiment of the present invention. The following describes the data analysis device of the energy storage and transmission micro-control module of the new energy electric vehicle in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a data analysis device for a new energy electric vehicle energy storage and transmission microcontroller module includes: The data acquisition module 201 is used to collect raw data from the energy storage and transmission microcontroller module of the new energy electric vehicle, and filter the raw data according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain a key data set; A synchronization verification module 202 is configured to perform cross-channel cache synchronization on the key data set and compare the timestamps and sequence number differences of adjacent data packets in each channel of the key data set to obtain a consistency verification report; The packet tracking module 203 is used to perform multi-dimensional packet tracking processing on the suspicious data packets marked in the consistency check report to obtain a list of out-of-order packets; The correlation analysis module 204 is used to perform correlation analysis on the energy storage transmission process of the new energy electric vehicle according to the out-of-sequence packet list, obtain the triggering conditions of the out-of-sequence packet list, and use the triggering conditions as the abnormal warning analysis results of the new energy electric vehicle.

[0047] In an embodiment of the present invention, the data analysis device of the energy storage and transmission microcontroller module of the new energy electric vehicle runs the data analysis method of the energy storage and transmission microcontroller module of the new energy electric vehicle. The data analysis device of the energy storage and transmission microcontroller module of the new energy electric vehicle performs time series acquisition and preliminary filtering on the vehicle's multi-channel raw data to obtain a key data set; then uses the bandwidth prediction feedback mechanism to perform cross-channel dynamic cache synchronization, and performs time series cross-checking on adjacent data packets to generate a consistency check report; then performs multi-dimensional packet tracking on the suspicious data packets in the report to form a list of out-of-order packets; finally, in the multi-level correlation analysis, the triggering conditions that lead to the generation of out-of-order packets are mined, and an abnormal warning result is output. The present invention overcomes the defect that sequence disorder caused by uneven bandwidth allocation and cache redundancy is difficult to detect in time by effectively retaining extreme fluctuation information under high load conditions and accurately identifying data packet disorder and retransmission, and can further enhance the data transmission reliability and fault warning capability of the energy storage system of the new energy electric vehicle.

[0048] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0049] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0050] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, 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 embodiments of the present invention.

Claims

1. A data analysis method for a new energy electric vehicle energy storage and transmission microcontroller module, characterized in that: The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module includes: Collecting raw data from the energy storage and transmission microcontroller module of the new energy electric vehicle, and filtering the raw data according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain a key data set; Performing cross-channel cache synchronization on the key data set, and comparing the timestamps and sequence number differences of adjacent data packets in each channel in the key data set to obtain a consistency check report; Perform multi-dimensional packet tracking on suspicious data packets marked in the consistency check report to obtain a list of out-of-order packets; According to the out-of-sequence packet list, the energy storage transmission process of the new energy electric vehicle is subjected to correlation analysis processing to obtain the triggering conditions of the out-of-sequence packet list, and the triggering conditions are used as abnormal warning analysis results of the new energy electric vehicle.

2. The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module according to claim 1 is characterized in that: The raw data of the energy storage and transmission microcontroller module of the new energy electric vehicle is collected, and the raw data is filtered according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain the key data sets including: Performing time series data collection on the power battery monitoring channel, temperature sensing channel, and inertial measurement unit channel in the energy storage and transmission microcontroller module to obtain raw data with channel characteristic labels; According to the load level and transmission priority of the new energy electric vehicle under different vehicle operating conditions, the continuous sampling records in the original data are segmented and split to obtain data to be filtered; According to the channel characteristics, transmission priority and vehicle operating conditions of the original data, a difference threshold determination process is performed on the abnormal extreme values ​​and repeated sampling points in the data to be filtered, and the abnormal extreme values ​​and repeated sampling points are deleted and merged respectively to obtain a key data set.

3. The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module according to claim 1 is characterized in that: The cross-channel cache synchronization of the key data set and the comparison of the timestamps and sequence number differences of adjacent data packets in each channel of the key data set to obtain a consistency check report include: Perform bandwidth prediction based on the historical bandwidth occupancy of each channel in the key data set to obtain a bandwidth demand prediction result in a future time window; Based on the bandwidth demand prediction result, dynamically assigning priorities to data packets of each channel in the key data set, and adjusting the cache space of each channel according to the assignment result to obtain a cross-channel cache allocation scheme; Performing cache synchronization processing on the key data set according to the cross-channel cache allocation scheme to obtain a synchronized multi-channel data stream; Performing difference calculation on the timestamps and sequence numbers of adjacent data packets in the multi-channel data stream to obtain a timestamp difference sequence and a sequence number difference sequence; According to the preset timestamp tolerance threshold and sequence number tolerance threshold, anomaly detection is performed on the timestamp difference sequence and sequence number difference sequence, suspicious data packets are marked, and a consistency check report is obtained.

4. The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module according to claim 1 is characterized in that: The multi-dimensional packet tracking process is performed on the suspicious data packets marked in the consistency check report to obtain the out-of-order packet list, including: Extract relevant data packets within a preset time window before and after the channel to which the suspicious data packet belongs according to the marking information in the consistency check report to form a local data packet sequence; Calculating the time interval and sequence number interval between the suspicious data packet and adjacent data packets based on the local data packet sequence to obtain an interval feature vector; Calling a preset multi-dimensional determination algorithm to classify the interval feature vector, determine the out-of-sequence type of the suspicious data packet, and obtain an out-of-sequence determination result; In combination with the vehicle operating conditions and channel load status, a context correlation analysis is performed on the out-of-sequence determination result to obtain an out-of-sequence packet list.

5. The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module according to claim 4 is characterized in that: The out-of-sequence packet list obtained by performing context correlation analysis on the out-of-sequence determination result in combination with the vehicle operating condition and the channel load state includes: Obtaining the channel load status of each channel in the corresponding time period, and performing time series segmentation processing on the vehicle operating condition, identifying the key time points of the operating condition change, and obtaining the operating condition change sequence; According to the working condition change sequence, each suspicious data packet in the out-of-sequence determination result is matched with the working condition to obtain a list of out-of-sequence candidates with working condition marks; Calculating a probability score of out-of-sequence occurrence for each data packet in the out-of-sequence candidate list according to the channel load status to obtain an out-of-sequence probability vector; The out-of-sequence probability vectors are screened according to a preset out-of-sequence determination threshold, and data packets with a probability higher than the threshold are marked as confirmed out-of-sequence, thereby obtaining an out-of-sequence packet list.

6. The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module according to claim 1 is characterized in that: The step of performing correlation analysis on the energy storage transmission process of the new energy electric vehicle according to the out-of-sequence packet list to obtain a triggering condition of the out-of-sequence packet list, and using the triggering condition as an abnormality warning analysis result of the new energy electric vehicle includes: Performing time cluster analysis on the out-of-order packets in the out-of-order packet list to obtain out-of-order event clusters; Extracting vehicle operating parameters within a corresponding time period based on the out-of-sequence event clusters, and performing time series analysis on the vehicle operating parameters to obtain a state transition sequence related to the out-of-sequence event; A preset association rule mining algorithm is called to extract association rules between out-of-sequence events and system state changes based on the state transition sequence, obtain trigger conditions for the out-of-sequence packet list, and use the trigger conditions as abnormal warning analysis results of the new energy electric vehicle.

7. The data analysis method of the new energy electric vehicle energy storage and transmission micro-control module according to claim 6 is characterized in that: The step of extracting vehicle operating parameters within a corresponding time period based on the out-of-sequence event clusters and performing time series analysis on the vehicle operating parameters to obtain a state transition sequence related to the out-of-sequence event includes: According to the clustering of the out-of-sequence events, extracting vehicle operating parameters of a preset time period before and after the corresponding time period from a vehicle status monitoring system; The vehicle operating parameters are segmented using a sliding time window method, and statistical features and frequency domain features are extracted within each time window to obtain a multidimensional state vector; Applying a principal component analysis algorithm to the multidimensional state vector to obtain a eigenvector after dimensionality reduction, and using a dynamic time warping algorithm to perform time sequence alignment on the eigenvector after dimensionality reduction to obtain a key parameter change sequence; According to a preset state transition rule library, the key parameter change sequence is mapped into discrete state identifiers and connected in chronological order to obtain a state transition sequence related to out-of-order events.

8. A data analysis device for a new energy electric vehicle energy storage and transmission microcontroller module, characterized in that: The data analysis device of the new energy electric vehicle energy storage and transmission micro-control module includes: A data acquisition module is used to collect raw data from the energy storage and transmission microcontroller module of the new energy electric vehicle, and filter the raw data according to the channel characteristics, transmission priority and vehicle operating conditions of the raw data to obtain a key data set; A synchronization verification module is used to synchronize the cross-channel cache of the key data set and compare the timestamps and sequence number differences of adjacent data packets in each channel of the key data set to obtain a consistency verification report; a packet tracking module, configured to perform multi-dimensional packet tracking processing on suspicious data packets marked in the consistency check report to obtain a list of out-of-order packets; The association analysis module is used to perform association analysis on the energy storage transmission process of the new energy electric vehicle according to the out-of-sequence packet list, obtain the triggering conditions of the out-of-sequence packet list, and use the triggering conditions as the abnormal warning analysis results of the new energy electric vehicle.

Images