A cloud platform-based battery remote diagnosis system

Abstract

The application relates to the technical field of battery diagnosis, in particular to a battery remote diagnosis system based on a cloud platform, which comprises a data cache collection module, a rate hopping extraction module, a region synchronization discrimination module, a combined behavior recognition module and a remote diagnosis decision module.In the application, the time sequence information constructed based on the incremental difference realizes fine capture of the voltage change trend, the voltage hopping feature is determined by combining the switching behavior of the continuous difference symbol, the hopping synchronization is recognized by linking the difference trend and the temperature change, the group pair relationship of the hopping time interval between regions is established, the region activity state is reflected by the hopping overlap behavior, the combined abnormal condition is discriminated by introducing the voltage current ratio and the temperature trend joint analysis in the active region, the critical behavior of the temperature rise abnormality superimposed on the electric parameter change is revealed, the dynamic sensitivity, the trend fusion degree and the region relevance of the battery abnormality recognition are effectively enhanced, and the recognition rate and the response accuracy of the complex fault mode of the battery are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery diagnosis, in particular to a battery remote diagnosis system based on a cloud platform. BACKGROUND

[0002] The technical field of battery diagnosis involves monitoring, analyzing and evaluating the running state, performance parameters and health status of various types of batteries, including real-time collection of key parameters such as battery voltage, current, temperature, internal resistance, life prediction, fault identification and remaining capacity estimation, etc. Through means such as sensor collection devices, battery management systems, data analysis platforms and communication networks, intelligent diagnosis and management of power batteries, energy storage batteries and consumer batteries are achieved, widely used in intelligent transportation, new energy equipment and industrial energy storage scenarios. Among them, the traditional battery remote diagnosis system refers to relying on local embedded collection units to record battery voltage, temperature, current and other operating parameters, and upload them to the control center, and then the local server judges the state of the battery through comparison analysis, voltage fluctuation analysis, historical trend comparison, etc.

[0003] In the traditional battery remote diagnosis process, the main method is to rely on local embedded devices for single-point data recording and centralized uploading, which lacks real-time guarantee. In the process of voltage and current mutation, it is difficult to continuously depict the evolution rhythm of parameters, and static comparison and trend comparison methods cannot identify regional or synchronous characteristics, resulting in the masking of group abnormal evolution trends in multi-battery systems. Especially in the early stage of fault induction, abnormal signals are easily masked by mean smoothing processing, causing alarm delay and fault positioning ambiguity, which affects the operation safety evaluation and operation and maintenance response efficiency in intelligent transportation and energy storage scenarios. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art and to provide a battery remote diagnosis system based on a cloud platform.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme: a battery remote diagnosis system based on a cloud platform comprises:

[0006] The data buffer collection module acquires battery basic parameters, constructs sampling data, performs difference operation on adjacent two sampling values of the same parameter and attaches original time stamp, stores according to battery number, generates time series incremental data;

[0007] The rate jump extraction module extracts the difference sign of any three consecutive time points and calculates the number of positive and negative direction switches according to the voltage difference data in the time series incremental data, records the voltage jump point, reads the temperature difference data at the jump point to judge whether the direction is changed from negative to positive, and generates a synchronous jump time sequence list;

[0008] The regional synchronization discrimination module calculates the time interval between the transition points of any two groups of batteries according to the synchronization transition time sequence list, marks the synchronization pairs, counts the number of synchronization pairs in the region, marks the region with a number greater than the synchronization base as an abnormally active region, and generates a regional synchronization transition identification record.

[0009] The combined behavior recognition module extracts the original voltage and current sequences within the corresponding time period based on the synchronous jump recognition records of the region, calculates the changing trend of the voltage-current ratio at each time point, filters the time points that simultaneously satisfy the ratio decline and temperature increase, and generates combined trend abnormal time ratio data.

[0010] As a further embodiment of the present invention, the time series incremental data includes parameter difference, original timestamp, and battery number; the synchronous jump time series table includes jump time, voltage difference direction, and temperature difference trend; the regional synchronous jump identification record includes regional number, number of synchronous pairs, and abnormal active status; and the combined trend abnormal time ratio data includes voltage-current ratio change rate, temperature rise trend characteristics, and jump point overlap information.

[0011] As a further aspect of the present invention, the data caching and acquisition module includes:

[0012] The parameter acquisition submodule acquires the basic parameters of battery voltage, current and temperature detected by the sensor, and periodically samples all parameters according to the basic sampling period and attaches the original timestamp to generate basic sampling data.

[0013] The incremental construction submodule reads the values ​​of the same parameter in two adjacent samples and their original timestamps based on the basic sampling data, calculates the difference between the two sample values, constructs the incremental record of the corresponding parameter, binds the difference result with the timestamp and classifies it according to the battery number to obtain the incremental data set.

[0014] The sequence generation submodule constructs a continuous data point sequence structure by sequentially concatenating the incremental data set according to the battery number, combining the differences of each parameter with the corresponding timestamp, and obtaining the time series incremental data.

[0015] As a further aspect of the present invention, the rate jump extraction module includes:

[0016] The symbol sequence construction submodule extracts the voltage difference values ​​of three consecutive time points corresponding to the battery number based on the voltage difference values ​​in the time series incremental data, determines the positive or negative value of the difference and extracts the corresponding symbol, constructs it into a symbol sequence, traverses the time axis to convert all data under the battery number that meet the triplet condition into a symbol sequence, records it into a symbol sequence buffer set, and obtains the voltage difference symbol sequence.

[0017] The jump point identification submodule counts the number of positive and negative sign switching in each triplet based on the voltage difference symbol sequence. When there are two direction switching, the center time point is extracted, and the absolute value of the voltage difference at the corresponding time is extracted from the incremental data. If it exceeds the voltage change threshold, the corresponding time point is marked as a jump point, and a set of marked jump time points is generated.

[0018] The timing filtering submodule extracts the temperature difference value of the time points before and after the jump point based on the set of marked jump time points, determines whether the temperature change direction has changed from negative to positive, and selects the time point that meets the condition as the synchronous jump point. All the jump point timestamps that meet the condition are arranged and combined in chronological order to obtain the synchronous jump time series list.

[0019] As a further aspect of the present invention, the regional synchronization discrimination module includes:

[0020] The jump point screening submodule obtains the synchronous jump time series list, groups it according to the area number to which the battery belongs, extracts the jump point timestamp sequence within a set analysis period for each area, calculates the timestamp interval of the jump points and performs a pairing index operation to generate a jump point time pair set within the area;

[0021] The synchronization pair determination submodule calculates the time interval between each pair of timestamps based on the set of jump point time pairs in the region, sets a synchronization time window, and records a synchronization pair if the time interval is less than the synchronization time window and the corresponding jump point belongs to a different battery number. It counts the total number of synchronization pairs in each region and calculates and outputs the region synchronization index.

[0022] The abnormal region marking submodule reads the number of synchronization pairs in each region according to the region synchronization index and compares it with the synchronization base number. If the number of synchronization pairs is greater than the synchronization base number, the corresponding region number is recorded as an abnormally active region, thus obtaining a region synchronization jump identification record.

[0023] As a further aspect of the present invention, the combined behavior recognition module includes:

[0024] The voltage-current ratio extraction submodule extracts the voltage and current sequences within the corresponding time period based on the battery numbers of the abnormally active areas marked in the regional synchronous jump identification record, obtains the ratio sequence at each sampling point, calculates the difference between two adjacent ratios, determines the trend according to the positive or negative direction of the difference, marks the direction of ratio change, and generates voltage-current ratio change trend data.

[0025] The temperature trend judgment submodule obtains the original temperature sequence corresponding to the battery number within the same time period based on the corresponding time period data in the voltage-current ratio change trend data, and sequentially judges whether the temperature values ​​of any three consecutive time points strictly increase. If the increase is true, it judges whether the temperature difference between the third time point and the first time point exceeds the temperature rise rate standard. If it exceeds, it is marked as meeting the abnormal temperature rise, and the abnormal temperature rise time point data is obtained.

[0026] The combined trend filtering submodule compares the voltage-current ratio change trend data with the abnormal temperature rise time point data to see if there is any overlap or adjacency between the time points. If the same time point or the maximum time error does not exceed a single sampling interval and simultaneously satisfies the ratio decline and temperature increase trends, it is included in the combined abnormal trend event. The combined trend abnormal time ratio value is calculated to reflect the proportion of abnormal combined trend events in the overall time interval, and the combined trend abnormal time ratio data is obtained.

[0027] As a further aspect of the present invention, the system further includes:

[0028] Based on the combined trend abnormal time ratio data, the remote diagnostic decision module marks batteries that exceed the abnormal time ratio threshold, sends the corresponding battery device number, time period and abnormal level back to the user terminal interface, and generates a cloud platform battery remote diagnostic record.

[0029] The cloud platform's remote battery diagnostic records include the battery number, abnormal time period, and functional safety level.

[0030] As a further aspect of the present invention, the remote diagnostic decision module includes:

[0031] The time proportion filtering submodule determines whether the value corresponding to each battery number is greater than the abnormal time proportion threshold based on the combined trend abnormal time proportion data, extracts the battery numbers that are greater than the abnormal time proportion threshold, and simultaneously obtains the corresponding sampling time period, filters the batteries that meet the set standards and the time range of abnormal behavior, and generates abnormal battery filtering result data.

[0032] The diagnostic data construction submodule obtains the selected battery number and corresponding time period based on the abnormal battery screening results data, reads the voltage, current, temperature and timestamp sequence in the corresponding segment in the original data record, combines and splices them according to the timestamp, performs a unified field standardization operation, sets different interval ranges to distinguish the abnormal performance level, extracts the abnormal intensity level category matched by each group of samples according to the interval correspondence, and obtains the remote diagnostic level data of the battery.

[0033] The anomaly level push submodule, based on the battery remote diagnostic level data, performs a matching operation between each group of battery numbers and their corresponding level values, retrieves the diagnostic time period and its corresponding diagnostic level, forms a structured dataset, writes it to the remote database platform, and simultaneously pushes it to the user terminal page refresh queue to obtain the cloud platform's battery remote diagnostic records.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0035] In this invention, time-series information constructed based on incremental differential is used to capture voltage change trends in a precise manner. The switching behavior of continuous differential symbols is used to determine voltage jump characteristics. The jump synchronicity is identified by linking differential trends with temperature changes. A pairing relationship of jump time intervals between regions is established. The overlapping behavior of jumps reflects the activity status of regions. In active regions, the voltage-current ratio and temperature trend are introduced for joint analysis to identify combined abnormal situations. The critical behavior of temperature rise abnormality superimposed on electrical parameter changes is revealed. A remote diagnostic strategy with hierarchical identification and trend superposition capabilities is formed, which effectively enhances the dynamic sensitivity, trend fusion and regional correlation of battery anomaly identification, and improves the identification rate and response accuracy of complex battery fault modes. Attached Figure Description

[0036] Figure 1 This is a system flowchart of the present invention;

[0037] Figure 2 This is a flowchart of the data caching and acquisition module of the present invention;

[0038] Figure 3 This is a flowchart of the rate jump extraction module of the present invention;

[0039] Figure 4 This is a flowchart of the regional synchronization discrimination module of the present invention;

[0040] Figure 5 This is a flowchart of the combined behavior recognition module of the present invention;

[0041] Figure 6 This is a flowchart of the remote diagnostic decision module of the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0043] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0044] Please see Figure 1 A cloud-based remote battery diagnostic system includes:

[0045] The data caching and acquisition module acquires the battery's basic parameters such as voltage, current, and temperature through sensors, constructs sampling data according to fixed time intervals (using the 1-second basic sampling period in the SAEJ1939-11 standard), performs difference calculations on two adjacent sampling values ​​of the same parameter and attaches the original timestamp, stores the data according to the battery number, and generates time series incremental data.

[0046] The rate jump extraction module extracts the differential sign of any three consecutive time points from the voltage differential data in the time series incremental data and calculates the number of positive and negative direction switching. When there are two sign switching and the current absolute value of the difference is greater than the voltage change threshold (≥5% / second of the rated voltage), the corresponding time is recorded as the voltage jump point. At the same time, the temperature differential data is read at the jump point to determine whether the direction has changed from negative to positive. Time points that meet the conditions are filtered and a synchronous jump time series table is generated.

[0047] The regional synchronization discrimination module classifies and extracts the jump points within a set period according to the synchronization jump time sequence list by region number, calculates the jump point time interval between any two groups of batteries, and marks them as a synchronization pair when the time interval is less than the synchronization time window within the group (≤200ms according to GB / T31467.3-2015). The module counts the number of synchronization pairs in the region and compares it with the synchronization base. When the number is greater than the synchronization base, the corresponding region is marked as an abnormally active region, and a regional synchronization jump identification record is generated.

[0048] The combined behavior recognition module extracts the original voltage and current sequences for the corresponding time period based on the battery numbers marked as abnormally active areas in the regional synchronous jump recognition record, calculates the changing trend of the voltage-current ratio at each time point, determines whether the temperature sequence increases for three consecutive time points, filters time points that simultaneously satisfy the ratio decline and temperature increase (satisfying the abnormal temperature rise rate defined by IEC62133-2:2017: ≥1℃ / min), and generates combined trend abnormal time ratio data by combining the jump point overlap situation.

[0049] The remote diagnostic decision module, based on the combined trend abnormal time ratio data, marks batteries that exceed the abnormal time ratio threshold (≥30% as set in SAEJ2931-2020), and transmits the corresponding battery device number, time period, and abnormal level (using ISO26262-2018 functional safety level (ASIL classification)) back to the user terminal interface, and generates a cloud platform battery remote diagnostic record.

[0050] The time-series incremental data includes parameter differences, original timestamps, and battery numbers. The synchronous jump time series table includes jump times, voltage difference directions, and temperature difference trends. The regional synchronous jump identification records include regional numbers, number of synchronous pairs, and abnormal activity status. The combined trend abnormal time ratio data includes voltage-current ratio change rate, temperature rise trend characteristics, and jump point overlap information. The cloud platform battery remote diagnostic records include battery numbers, abnormal time periods, and functional safety levels.

[0051] Please see Figure 2 The data caching and acquisition module includes:

[0052] The parameter acquisition submodule acquires the basic parameters of battery voltage, current and temperature detected by the sensor, and periodically samples all parameters according to the basic sampling period and attaches the original timestamp to generate basic sampling data.

[0053] To obtain basic parameters such as battery voltage, current, and temperature detected by sensors, a temperature acquisition unit consisting of a voltage detector, a current sensor, and a thermistor should first be configured on each module of the power battery pack. All sensors are connected to the sampling control system via the CAN bus protocol, and sampling is performed once per second according to the SAE J1939-11 standard. For example, for a battery cell with the serial number B101, at 12:00:00, the voltage is 3.65V, the current is 1.25A, and the temperature is 35.2℃. At 12:00:01, a second sampling yields a voltage of 3.67V, a current of 1.28A, and a temperature of 35.5℃. This type of raw parameter... Numerical values ​​are synchronously bound to a UNIX timestamp and written to a buffer. The data structure includes fields such as sampling sequence number, time, parameter type, parameter value, and battery number. Sampled values ​​are uniformly represented in float format by the ADC conversion module. Each sensor value needs to be verified according to the module's accuracy range; for example, the voltage acquisition accuracy is set to 0.01V, current to 0.01A, and temperature to 0.1℃. The sampling buffer module generates an independent sampling record row for each second of data, indexed by a combination of record number and battery ID. The record format for each data entry is: {sampling sequence number, sampling time, voltage value, current value, temperature value}. The data structure is shown below:

[0054] Table 1 Initial Data Table of Sampling Parameters

[0055]

[0056] As shown in Table 1, the original values ​​of each parameter recorded in the two consecutive sampling points of battery number B101 were all collected under the standard cycle. This set of data is used as the original sampling input, and the sampling time interval is marked as 1 second. The data point structure is complete and continuous, and the basic sampling data is finally obtained.

[0057] The incremental construction submodule reads the values ​​of the same parameter in two adjacent samples and their original timestamps based on the basic sampling data, calculates the difference between the two sample values, constructs the incremental record of the corresponding parameter, binds the difference result with the timestamp and classifies it by battery number to obtain the incremental data set;

[0058] Based on the basic sampling data, the values ​​of the same parameter in two adjacent samplings and their original timestamps are retrieved. Differences are extracted for voltage, current, and temperature. First, the time interval Δt = 1 second is set. Based on the two voltage sampling values ​​of 3.65V and 3.67V in Table 1, the voltage difference ΔU = 0.02V; the current difference ΔI = 1.28A. 1.25A = 0.03A; Temperature difference ΔT = 35.5℃ 35.2℃ = 0.3℃. The above difference items are categorized and bound. The difference, together with the timestamp, parameter type, and battery number, are used to construct a data tuple. For example, the voltage change item is {voltage, 12:00:01, 0.02, B101}. Parameter difference filtering thresholds are set to remove interfering items. For example, the voltage change threshold is set to 0.03V, the current to 0.05A, and the temperature to 0.5℃. Currently, none of the three differences exceed the threshold, so they are included in the valid incremental data. In the cache structure, the difference sequence of different parameters is hashed to the memory index table according to the battery number. Different numbered units are distinguished by B101, B102, etc. For example, the voltage difference sequence under the number B101 is [0.02], the current is [0.03], and the temperature is [0.3]. All data are sorted in ascending order by timestamp and written into the set structure, and finally the incremental data set is obtained.

[0059] The sequence generation submodule concatenates the incremental data set according to the battery number, combines the differences of each parameter with the corresponding timestamp, constructs a continuous data point sequence structure, and obtains time series incremental data.

[0060] Based on the incremental data set, the data is sequentially concatenated according to the battery number. Incremental records for voltage, current, and temperature are read separately. Using an index structure, each difference item is combined according to parameter type. For example, the voltage increment sequence for battery B101 is {12:00:01, 0.02}. If subsequent sampling yields {12:00:02, 0.01}, it is concatenated into the complete sequence [{12:00:01, 0.02}, {12:00:02, 0.01}]. The current and temperature sequences are constructed similarly. The data format uses a list or array structure for subsequent program parsing. Each parameter difference sequence contains the trajectory of changes at continuous time points. Separate data blocks are created for different parameters, using label fields to identify parameter type and number. Finally, a complete set of time increment sequences for each battery number is formed. This set will serve as crucial input data for determining the dynamic operating status of individual batteries in subsequent system processing, ultimately obtaining the time series incremental data.

[0061] Please see Figure 3 The rate jump extraction module includes:

[0062] The symbol sequence construction submodule extracts the voltage difference value of three consecutive time points corresponding to the battery number based on the voltage difference value in the time series incremental data, determines the positive or negative value of the difference and extracts the corresponding symbol, constructs it into a symbol sequence, traverses the time axis and converts all data under the battery number that meet the triplet condition into a symbol sequence, records it into a symbol sequence buffer set, and obtains the voltage difference symbol sequence.

[0063] Based on the voltage difference values ​​in the time series incremental data, a voltage difference sequence is first established for each battery number. A sliding triplet is constructed for the voltage differences at three consecutive time points in each group, and the sign of each difference is determined, using the battery number as the reference. Taking a single battery cell as an example, the sampling times are as follows: , , The corresponding voltage difference is , , After symbol extraction, they are converted into "−", "+", and "−" respectively, and then assembled into a symbol sequence. Each symbol triplet needs to record its corresponding timestamp and battery number information and store it in the symbol sequence buffer list. At the same time, the corresponding original difference sequence is retained for subsequent use. The sampling results are shown in the table:

[0064] Table 2 Examples of Voltage Differential Symbol Construction

[0065]

[0066] As shown in Table 2, the voltage difference data has been symbolically encoded in chronological order and a triplet structure has been established to generate a voltage difference symbol sequence.

[0067] The jump point identification submodule is based on the voltage differential symbol sequence. It counts the number of positive and negative sign switching in each triplet. When there are two direction switching, it extracts the center time point and extracts the absolute value of the voltage difference at the corresponding time from the incremental data. If it exceeds the voltage change threshold, the corresponding time point is marked as a jump point and a set of marked jump time points is generated.

[0068] Based on the voltage differential symbol sequence, for each triplet The number of direction switching events is counted; if two positive and negative transitions occur, the intermediate time point is extracted. This corresponds to a candidate point for the voltage jump, and then the absolute voltage difference at that time point is read. Let the rated voltage be The mutation threshold is calculated according to the rules. ,like Then mark this time point as the jump time point, and continue to extract temperature difference data. Let the temperature difference before and after this time point be... and Determine whether it changes from negative to positive. If the condition is met, retain the point as a valid transition record and obtain the set of marked transition times.

[0069] The timing filtering submodule extracts the temperature difference value of the time points before and after the jump point based on the set of marked jump time points, determines whether the temperature change direction has changed from negative to positive, and selects the time point that meets the condition as the synchronous jump point. All the jump point timestamps that meet the condition are arranged and combined in chronological order to obtain the synchronous jump time series list.

[0070] Extract the timestamp of each transition point from the set of marked transition moments. And read the point and its previous time point. Corresponding temperature difference and Execute the judgment condition and If this condition is met, then mark the current time point as a synchronous temperature rise type jump point, for example, at the point numbered... In the battery, if hour, , If the condition is met, it is recorded as a synchronous transition point. All transition points that meet the conditions will be sorted in ascending order of time and added to the result array. The array fields include... Finally, the synchronization transition sequence list is obtained.

[0071] Please see Figure 4The regional synchronization discrimination module includes:

[0072] The jump point filtering submodule obtains the synchronous jump time series list, groups it according to the area number to which the battery belongs, extracts the jump point timestamp sequence within the set analysis period for each area, calculates the timestamp interval of the jump points and performs pairing indexing operation to generate a jump point time pair set within the area;

[0073] After obtaining the synchronous jump time series table, first read the "Battery Number" and its corresponding "Region Number" fields from the jump records. Group all jump points by region number. For example, Table 3 shows the battery numbers and jump time data corresponding to regions A1 and A2. In region A1, batteries numbered B101, B102, and B103 experienced jump events at 12:00:00, 12:00:00.15, and 12:00:00.35, respectively. Set the analysis period. Then, all valid transition point timestamps need to be extracted within 60 seconds, and batteries in the same area need to be paired up to generate transition point time pairs. For example, in area A1, the pairings are: , , Next, the timestamp field is read, and the three time intervals are calculated respectively to obtain... , , The records are then organized into structured records according to region numbering, as shown in the table below:

[0074] Table 3. Examples of time jump points within the region

[0075]

[0076] As shown in Table 3, the time pairing of transition points within the region has been completed, and the time difference has been marked and recorded to generate a set of time pairs of transition points within the region.

[0077] The synchronization determination submodule calculates the time interval between each pair of timestamps based on the set of time pairs of transition points within the region, and sets the synchronization time window accordingly. If the time interval is less than the synchronization time window and the corresponding jump point belongs to a different battery number, it is recorded as a synchronization pair. The total number of synchronization pairs in each region is counted using the formula:

[0078] ;

[0079] The operation output area synchronization index, among which... Indicates regional synchronization indicators. For the first For the time interval of the transition point, This serves as a reference value for the synchronization time window. For the corresponding physical distance of the battery (unit: meters). As a normalized distance reference, For the total number of regions, For the region Number of transition points in the middle;

[0080] Based on the set of time pairs of transition points within the region, set a synchronization time window threshold. Each time pair is checked individually to determine whether it meets the synchronization criteria. If the battery numbers are inconsistent, they are recorded as a synchronization pair, for example, in area A1, the time pair... The time interval of 0.15 seconds meets the condition. A time interval of 0.20 seconds is also acceptable. The timeout of 0.35 seconds is not counted. The number of synchronization pairs in area A1 is 2, recorded as follows. Further calculate the regional synchronization index.

[0081] Set the total number of regions In region A1, the two pairs of time pairs that meet the conditions are:

[0082] ;

[0083] ;

[0084] in, , The corresponding product is ;

[0085] , The corresponding product is ;

[0086] The synchronization index for region A1 is:

[0087] ;

[0088] This value is the synchronization index for region A1, generating the region synchronization index.

[0089] The regional synchronization index is used to quantify the synchronous activity of jump events in the time and space dimensions within a certain battery physical area. Its value comprehensively reflects the density of jump points occurring within a short time window and the spatial proximity between them. The larger the value, the more frequent the phenomenon of voltage and temperature sudden changes occurring simultaneously or approximately simultaneously among different batteries in the area within a very short time. Furthermore, the smaller the physical distance between these jump points, the more likely they are to exhibit a concentrated behavior trend influenced by the same triggering factor, and the higher the anomalous clustering. Therefore, this index can serve as a core judgment basis for identifying potential systemic fluctuations or local anomalies within a battery area.

[0090] The formula reflects the coupling effect of time synchronization and spatial proximity. Firstly, it is derived by setting the time interval for each set of transition points. Synchronize with the set time window The ratio is normalized to reflect whether the jump is within a specified time window. This part uses absolute values ​​to eliminate the influence of the order of events while keeping the value non-negative. Then, the normalized time difference is compared with the physical distance corresponding to the jump point. Multiply by the reciprocal, and the physical distance is factored through a normalization factor. Transforming it into a dimensionless form, it emphasizes that the greater the spatial proximity between transition points, the greater their synchronization contribution. The product result is a joint index of temporal synchronization and spatial concentration. The synchronization coupling value of the entire region is obtained by summing all effective transition point combinations within each region. To reflect the overall comparison between regions, the sum of all regions is normalized and divided by the number of regions. The final output is a unit-independent regional synchronization index value, used to identify the strength of synchronization activity of jump events within a region. This structure correlates and fuses temporal compression and spatial density through multiplication, and performs global accumulation through addition, expressing the composite characteristics of synchronization events under the overall spatial distribution.

[0091] The abnormal region marking submodule reads the number of synchronization pairs in each region based on the region synchronization index and compares it with the synchronization base number. If the number of synchronization pairs is greater than the synchronization base number, the corresponding region number is recorded as an abnormally active region, thus obtaining the region synchronization jump identification record.

[0092] Based on the regional synchronization metrics, extract the number of synchronization pairs for each region and perform threshold judgment to set a synchronization baseline. The baseline was determined experimentally. Ten 60-second periodic segments were sampled during the experiment, and the frequency of transitions and the number of synchronization pairs within each segment were compared. The results showed that the average number of synchronization pairs in high-activity areas was 1.8, and in ordinary areas it was 1.2. A comprehensive boundary value of 1.5 was set, and the number of synchronization pairs in the current area A1 was 2, satisfying the requirement. Therefore, region A1 is recorded as an abnormally active region, and relevant fields are compiled and written into the results record table, including region number, number of synchronization pairs, average time interval, synchronization indicators, etc. The result record format is as follows:

[0093] Area code: A1;

[0094] Synchronous logarithm: 2;

[0095] Synchronous indicator: 1.375;

[0096] Exception flag: True;

[0097] The final result is the regional synchronous jump recognition record.

[0098] Please see Figure 5 The combined behavior recognition module includes:

[0099] The voltage-current ratio extraction submodule extracts the voltage and current sequences within the corresponding time period based on the battery numbers of abnormally active areas marked in the regional synchronous jump identification record, obtains the ratio sequence at each sampling point, calculates the difference between two adjacent ratios, determines the trend according to the positive or negative direction of the difference, marks the direction of ratio change, and generates voltage-current ratio change trend data.

[0100] Based on the abnormally active area battery numbers marked in the regional synchronous jump identification record, the voltage and current sample values ​​of that battery number are read from the original data table. The voltage value at the same timestamp is divided by the current value to obtain the voltage-current ratio. For example, at time point... The corresponding voltage is 3.70V and the current is 1.20A, then the corresponding ratio is... The ratio at subsequent time points is calculated in the same manner. , , Then, the ratio sequence is arranged in chronological order, and the difference operation is performed on the ratio at any two adjacent time points, such as... , , The trend of the ratio change is determined by judging the sign of the difference. When the difference is less than 0, it is recorded as "sloping down". In this way, the voltage-current ratio change trend sequence between each sampling point is generated, and the voltage-current ratio change trend data is obtained.

[0101] Table 4. Calculation of Battery Voltage, Current, and Ratio

[0102]

[0103] As shown in Table 4, the calculation of the voltage-current ratio and the determination of its changing trend at continuous time points have been completed.

[0104] The temperature trend judgment submodule obtains the original temperature sequence corresponding to the battery number within the same time period based on the voltage-current ratio change trend data of the corresponding time period data, and sequentially judges whether the temperature values ​​of any three consecutive time points strictly increase. If the increase is true, it judges whether the temperature difference between the third time point and the first time point exceeds the temperature rise rate standard. If it exceeds, it is marked as meeting the abnormal temperature rise, and the abnormal temperature rise time point data is obtained.

[0105] Based on the time period corresponding to the voltage-current ratio change trend data, temperature sampling sequences within the same time period are obtained. After arranging them in chronological order, each set of three consecutive temperature values ​​is checked to see if they strictly increase. For example, if the temperature sequence is 38.5℃, 39.2℃, 40.1℃, and 41.0℃, then a three-point group is formed. The temperature must satisfy a strict increasing condition. Next, determine the rate of temperature rise over the time span between these three points. The calculation method is as follows: , The rate of temperature rise is If the temperature / min exceeds the threshold defined in IEC 62133-2:2017 (1℃ / min), then record the time point. If the abnormal temperature rise condition is met, continue to judge the three-point group. have to , The rate is 2.7℃ / min, which also meets the requirements. Record the time points. These are the outliers, and the data for the time points of abnormal temperature rise are obtained.

[0106] Table 5. Battery Temperature Sequence and Rate of Increase Determination

[0107]

[0108] As shown in Table 5, time points t3 and t4 are marked as abnormal temperature rise time points.

[0109] The combined trend screening submodule compares the voltage-current ratio change trend data with the abnormal temperature rise time point data to check for time overlap or adjacency. If the same time point or the maximum time error does not exceed a single sampling interval and simultaneously satisfies both the ratio decline and temperature increase trends, it is counted as a combined abnormal trend event, using the formula:

[0110] ;

[0111] The calculation obtains the proportion of abnormal time periods in the combined trend, reflecting the percentage of abnormal combined trend events within the overall time interval. This yields the combined trend abnormal time proportion data. This represents the proportion of time when the combined trend is abnormal, and the unit is dimensionless. Indicates the first The change in the ratio of voltage to current at each time point compared to the previous time point. This is the voltage-to-current ratio at the previous time point. Indicates the first Temperature changes at specific points in time, in degrees Celsius. The time interval between this point in time and the previous point in time, in seconds. and The first Current and voltage values ​​at specific time points, in amperes and volts. This represents the ratio of current to voltage; its squared value reflects the fluctuation of the electrical load at that moment. Indicates in Perform a summation operation at each time point that satisfies the combined trend condition;

[0112] Based on the voltage-current ratio change trend data and the abnormal temperature rise time point data, determine whether there is a combined trend, and search Tables 4 and 5 for time points that simultaneously show a decline in the ratio and abnormal temperature rise to locate the time point. and If both conditions are met, extract the required data and substitute them into the formula.

[0113] Time point :

[0114] , ;

[0115] , ;

[0116] ;

[0117] product: ;

[0118] Time point :

[0119] , ;

[0120] , ;

[0121] ;

[0122] product: ;

[0123] Calculate the average:

[0124] ;

[0125] The results indicate that the abnormal time ratio of the combined trend is 0.004695, which is lower than the identification benchmark value of 0.0075. The combined trend in the current period does not constitute an abnormal trigger condition, and the abnormal time ratio data of the combined trend is finally obtained.

[0126] The combined trend anomaly time ratio value is used to measure the intensity of the time proportion during which the battery simultaneously experiences a decline in the voltage-to-current ratio and a continuous rise in temperature within the sampling period. Its value comprehensively reflects the coupling degree between the electrical change rate, the temperature rise amplitude, and the voltage and current load characteristics. The higher the value of this index, the more time points that simultaneously meet the conditions of sudden change in voltage ratio and increase in temperature, indicating that the battery has a stronger degree of combined stress in that period. This value can serve as a core reference for the quantitative identification of multidimensional abnormal trend behavior, and can be used to judge whether there is a linkage risk in the battery's operating state from both the aspects of persistence and intensity.

[0127] The formula first calculates the change in voltage-current ratio at each time point. The ratio to the previous moment After normalization, a dimensionless index of relative decline intensity is formed. It is used to measure the degree of instantaneous change in an electrical system, and also includes the amount of temperature change. With sampling time interval Divide to obtain the rate of temperature rise This is used to quantify the intensity of thermal abrupt changes. The two are coupled multiplicatively to reflect the degree of synchronization of the combined electrothermal fluctuations, and then the current-voltage ratio is introduced. Constructing nested forms of squared terms This nested structure, used to enhance the correction of trend strength by electrical load characteristics, increases the weighted response when the ratio deviates from the stable value. The three products ultimately reflect the coupling degree of electrical ratio jump, heat rise rate, and load state. The summation is performed on all time points that meet the conditions, and then divided by the total number of time points that meet the conditions. Obtain the average index This refers to the combined trend anomaly time ratio value. This structure integrates multiple exponential coupling effects through addition, constructs a common reinforcement mechanism of influencing factors through multiplication, and responds to load fluctuations through square root nonlinearity. The overall design is used to capture compound anomaly trends.

[0128] Please see Figure 6 The remote diagnostic decision-making module includes:

[0129] The time proportion filtering submodule is based on the combined trend abnormal time proportion data. It determines whether the value corresponding to each battery number is greater than the abnormal time proportion threshold, extracts the battery numbers that are greater than the abnormal time proportion threshold, and simultaneously obtains the corresponding sampling time period. It filters batteries that meet the set criteria and the time range of abnormal behavior, and generates abnormal battery filtering result data.

[0130] Based on the combined trend anomaly time proportion value, firstly, the proportion value of each battery group and its corresponding proportion value is obtained. The combined trend anomaly time proportion value is used to reflect the proportion of time during which the voltage-current ratio decreases and the temperature increases simultaneously within the observation period for each battery. The original proportion value data is obtained from the sampling result table, and a conditional judgment operation is performed on each proportion value. The judgment method is to compare the proportion value with the anomaly time proportion threshold. This threshold is set to 0.30 according to the SAE J2931-2020 standard. If the proportion value of the battery is greater than this threshold, it is judged as an abnormal battery. When performing this judgment, all sample batteries are traversed sequentially. The system assigns numbers to each battery and compares their proportion values. When the result is "yes," the corresponding number is extracted and its anomaly determination flag is recorded. Simultaneously, the corresponding time period field in the original data for that number is read. The battery numbers determined to be abnormal are combined with their time periods to form a preliminary abnormal device record group, which serves as the basis for subsequent diagnostics. For example, in the table, number B002 has a proportion value of 0.34, which is greater than the threshold and is marked as abnormal, while number B004 has a proportion value of 0.29, which is less than the threshold and is not included in the filtering result set. Finally, the battery numbers that meet the conditions and their time period information are retained to generate abnormal battery filtering result data.

[0131] Table 6. Results of Abnormal Battery Screening

[0132]

[0133] As shown in Table 6, three sets of battery numbers that meet the anomaly criteria and their corresponding sampling time periods have been selected from the sampling set.

[0134] The diagnostic data construction submodule obtains the selected battery number and corresponding time period based on the abnormal battery screening results data, reads the voltage, current, temperature and timestamp sequence in the corresponding segment in the original data record, combines and splices them according to the timestamp, performs a unified field standardization operation, sets different interval ranges to distinguish the abnormal performance level, extracts the abnormal intensity level category matched by each group of samples according to the interval correspondence, and obtains the remote diagnostic level data of the battery.

[0135] Based on the abnormal battery screening results, the device information and sampling time periods for devices numbered B002, B003, and B005 were obtained. For each number, field extraction was performed within the corresponding time period, including timestamp, voltage, current, temperature, SoC capacity, unit number, and device ID. All fields were concatenated and arranged according to the timestamp, unifying multiple fields to the same time series. Data standardization was then performed, retaining two decimal places (V) for voltage, two decimal places (A) for current, one decimal place (°C) for temperature, and normalizing the SoC to a real number between 0 and 1. The interval is then used to insert the combined trend anomaly time ratio value as a field into a standardized field table. Each record is then divided into grade intervals based on the ratio value. For example, values ​​between 0.30 and 0.35 are mapped to grade "L1", between 0.35 and 0.40 to "L2", and values ​​greater than 0.40 to "L3". For instance, battery number B003 has a ratio value of 0.41, falling into the "L3" interval, so its grade field is assigned "L3". Finally, a structure merging operation is performed on all fields to construct a structured multi-field diagnostic record set, forming a complete, time-series consistent, and graded device-side diagnostic data structure, resulting in remote battery diagnostic grade data.

[0136] Table 7 Sample Table of Battery Remote Diagnostic Levels (Excerpt B003)

[0137]

[0138] As shown in Table 7, a diagnostic level field has been constructed and a standardized structure has been formed for the abnormal battery data with the number B003.

[0139] The anomaly level push submodule, based on the remote battery diagnostic level data, matches each group of battery numbers with the corresponding level values, retrieves the diagnostic time period and the corresponding diagnostic level, forms a structured dataset, writes it to the remote database platform, and simultaneously pushes it to the user terminal page refresh queue to obtain the cloud platform's remote battery diagnostic records.

[0140] Based on the remote battery diagnostic level data, the diagnostic level field is read record by record. For each record, the functional safety level is matched according to the value of the level field. The local level fields L1, L2, and L3 are mapped to the ASILA, ASILAB, and ASILD levels in the functional safety standard, respectively. The mapping operation adopts an interval conversion method. If the level is "L1", it corresponds to ASILA; if the level is "L2", it corresponds to ASILAB; and if the level is "L3", it corresponds to ASILD. After the mapping is completed, the battery number, sampling time period, and functional safety level field in each group of records are extracted to form a structural unit. A structured information set indexed by the battery is constructed and written into the remote cloud platform diagnostic database table. Each record contains three core field values. A set of diagnostic record objects is generated by filling the structure. Then, each unit in the set is synchronously sent to the user terminal. The front-end interface data refresh interface is called to make the pushed information visually displayed on the terminal page and obtain the cloud platform battery remote diagnostic records.

[0141] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A cloud-based remote battery diagnostic system, characterized in that, The system includes: The data caching and acquisition module obtains basic battery parameters, constructs sampling data, performs difference calculations on two adjacent sampling values ​​of the same parameter and attaches the original timestamp, stores it according to the battery number, and generates time series incremental data. The rate jump extraction module extracts the differential sign of any three consecutive time points based on the voltage differential data in the time series incremental data and calculates the number of positive and negative direction switching, records the voltage jump point, reads the temperature differential data at the jump point to determine whether the direction has changed from negative to positive, and generates a synchronous jump time series table. The regional synchronization discrimination module calculates the time interval between the transition points of any two groups of batteries according to the synchronization transition time sequence list, marks the synchronization pairs, counts the number of synchronization pairs in the region, marks the region with a number greater than the synchronization base as an abnormally active region, and generates a regional synchronization transition identification record. The combined behavior recognition module extracts the original voltage and current sequences within the corresponding time period based on the regional synchronous jump recognition records, calculates the changing trend of the voltage-current ratio at each time point, filters the time points that simultaneously satisfy the ratio decline and temperature increase, and generates combined trend abnormal time ratio data. The rate jump extraction module includes: The symbol sequence construction submodule extracts the voltage difference values ​​of three consecutive time points corresponding to the battery number based on the voltage difference values ​​in the time series incremental data, determines the positive or negative value of the difference and extracts the corresponding symbol, constructs it into a symbol sequence, traverses the time axis to convert all data under the battery number that meet the triplet condition into a symbol sequence, records it into a symbol sequence buffer set, and obtains the voltage difference symbol sequence. The jump point identification submodule counts the number of positive and negative sign switching in each triplet based on the voltage difference symbol sequence. When there are two direction switching, the center time point is extracted, and the absolute value of the voltage difference at the corresponding time is extracted from the incremental data. If it exceeds the voltage change threshold, the corresponding time point is marked as a jump point, and a set of marked jump time points is generated. The timing filtering submodule extracts the temperature difference value of the time points before and after the jump point based on the set of marked jump time points, determines whether the temperature change direction has changed from negative to positive, and selects the time point that meets the condition as the synchronous jump point. All the jump point timestamps that meet the condition are arranged and combined in chronological order to obtain the synchronous jump time series list.

2. The cloud-based remote battery diagnostic system according to claim 1, characterized in that, The time series incremental data includes parameter differences, original timestamps, and battery numbers. The synchronous jump time series table includes jump times, voltage difference directions, and temperature difference trends. The regional synchronous jump identification records include regional numbers, number of synchronous pairs, and abnormal activity status. The combined trend abnormal time ratio data includes voltage-current ratio change rate, temperature rise trend characteristics, and jump point overlap information.

3. The cloud-based remote battery diagnostic system according to claim 1, characterized in that, The data caching and acquisition module includes: The parameter acquisition submodule acquires the basic parameters of battery voltage, current and temperature detected by the sensor, and periodically samples all parameters according to the basic sampling period and attaches the original timestamp to generate basic sampling data. The incremental construction submodule reads the values ​​of the same parameter in two adjacent samples and their original timestamps based on the basic sampling data, calculates the difference between the two sample values, constructs the incremental record of the corresponding parameter, binds the difference result with the timestamp and classifies it according to the battery number to obtain the incremental data set. The sequence generation submodule constructs a continuous data point sequence structure by sequentially concatenating the incremental data set according to the battery number, combining the differences of each parameter with the corresponding timestamp, and obtaining the time series incremental data.

4. The cloud-based remote battery diagnostic system according to claim 1, characterized in that, The regional synchronization discrimination module includes: The jump point screening submodule obtains the synchronous jump time series list, groups it according to the area number to which the battery belongs, extracts the jump point timestamp sequence within a set analysis period for each area, calculates the timestamp interval of the jump points and performs a pairing index operation to generate a jump point time pair set within the area; The synchronization pair determination submodule calculates the time interval between each pair of timestamps based on the set of jump point time pairs in the region, sets a synchronization time window, and records a synchronization pair if the time interval is less than the synchronization time window and the corresponding jump point belongs to a different battery number. It counts the total number of synchronization pairs in each region and calculates and outputs the region synchronization index. The abnormal region marking submodule reads the number of synchronization pairs in each region according to the region synchronization index and compares it with the synchronization base number. If the number of synchronization pairs is greater than the synchronization base number, the corresponding region number is recorded as an abnormally active region, thus obtaining a region synchronization jump identification record.

5. The cloud-based remote battery diagnostic system according to claim 1, characterized in that, The combined behavior recognition module includes: The voltage-current ratio extraction submodule extracts the voltage and current sequences within the corresponding time period based on the battery numbers of the abnormally active areas marked in the regional synchronous jump identification record, obtains the ratio sequence at each sampling point, calculates the difference between two adjacent ratios, determines the trend according to the positive or negative direction of the difference, marks the direction of ratio change, and generates voltage-current ratio change trend data. The temperature trend judgment submodule obtains the original temperature sequence corresponding to the battery number within the same time period based on the corresponding time period data in the voltage-current ratio change trend data, and sequentially judges whether the temperature values ​​of any three consecutive time points strictly increase. If the increase is true, it judges whether the temperature difference between the third time point and the first time point exceeds the temperature rise rate standard. If it exceeds, it is marked as meeting the abnormal temperature rise, and the abnormal temperature rise time point data is obtained. The combined trend filtering submodule compares the voltage-current ratio change trend data with the abnormal temperature rise time point data to see if there is any overlap or adjacency between the time points. If the same time point or the maximum time error does not exceed a single sampling interval and simultaneously satisfies the ratio decline and temperature increase trends, it is included in the combined abnormal trend event. The combined trend abnormal time ratio value is calculated to reflect the proportion of abnormal combined trend events in the overall time interval, and the combined trend abnormal time ratio data is obtained.

6. The cloud-based remote battery diagnostic system according to claim 1, characterized in that, The system also includes: Based on the combined trend abnormal time ratio data, the remote diagnostic decision module marks batteries that exceed the abnormal time ratio threshold, sends the corresponding battery device number, time period and abnormal level back to the user terminal interface, and generates a cloud platform battery remote diagnostic record. The cloud platform's remote battery diagnostic records include the battery number, abnormal time period, and functional safety level.

7. The cloud-based remote battery diagnostic system according to claim 6, characterized in that, The remote diagnostic decision module includes: The time proportion filtering submodule determines whether the value corresponding to each battery number is greater than the abnormal time proportion threshold based on the combined trend abnormal time proportion data, extracts the battery numbers that are greater than the abnormal time proportion threshold, and simultaneously obtains the corresponding sampling time period, filters the batteries that meet the set standards and the time range of abnormal behavior, and generates abnormal battery filtering result data. The diagnostic data construction submodule obtains the selected battery number and corresponding time period based on the abnormal battery screening results data, reads the voltage, current, temperature and timestamp sequence in the corresponding segment in the original data record, combines and splices them according to the timestamp, performs a unified field standardization operation, sets different interval ranges to distinguish the abnormal performance level, extracts the abnormal intensity level category matched by each group of samples according to the interval correspondence, and obtains the remote diagnostic level data of the battery. The anomaly level push submodule, based on the battery remote diagnostic level data, performs a matching operation between each group of battery numbers and their corresponding level values, retrieves the diagnostic time period and its corresponding diagnostic level, forms a structured dataset, writes it to the remote database platform, and simultaneously pushes it to the user terminal page refresh queue to obtain the cloud platform's battery remote diagnostic records.

Images