Vehicle battery monitoring method and device

Abstract

The application discloses a kind of vehicle-mounted battery monitoring method and device, the method includes: based on the single battery cycle use data in a certain period of time in advance, respectively establish first prediction model and second prediction model;Real-time monitoring vehicle-mounted battery data, and respectively according to the first prediction model and second prediction model are predicted, obtain first group SOH value and second group SOH value;According to the first group SOH value, generate the first data sequence of cumulative charge-predicted SOH value, and according to the second group SOH value, generate the second data sequence of cumulative charge-predicted SOH value;According to the first data sequence and the second data sequence, determine whether cell is abnormal, and / or whether cell attenuation trend is abnormal.Utilize the present application, can accurately judge cell abnormality and attenuation trend abnormality, to find problem cell early, reduce the security risk due to cell abnormality and consistency decline.

Description

Technical Field

[0001] This invention relates to the field of vehicle-mounted power lithium batteries, and specifically to a method and device for monitoring vehicle-mounted batteries. Background Technology

[0002] Lithium-ion batteries are widely used in the new energy vehicle industry due to their advantages such as high energy density, rechargeability, long cycle life, and no self-discharge. Typically, battery strings in vehicles consist of individual cells connected in series. During repeated charge-discharge cycles, battery performance degrades. Some individual cells degrade significantly faster than their connected counterparts. Reduced battery consistency can lead to overcharging or over-discharging of individual cells during charging and discharging, reducing the lifespan of the battery pack. Therefore, accurate prediction of the State of Health (SOH) of individual cells and assessment of cell degradation trends are crucial for ensuring battery safety and extending battery life.

[0003] In recent years, methods for predicting and estimating the state of harmonics (SOH) of batteries have been extensively studied. Common analytical methods include equivalent circuit model prediction and data-driven SOH prediction models. Equivalent circuit models are widely used, but these methods require the construction of an equivalent model, resulting in high computational costs, low timeliness, and inability to provide rapid online evaluation. Data-driven SOH prediction methods can reduce the complexity of model building and improve prediction efficiency; however, these methods require a large amount of data as a foundation, and the accuracy of the model is directly related to the quality of the data.

[0004] Existing data-driven battery SOH prediction methods can screen out individual cells with low SOH values ​​based on prediction results, but they cannot identify individual cells in series batteries with significant degradation trends to provide early warnings. These cells with significant degradation trends are difficult to screen in the early stages, potentially leading to safety hazards. Furthermore, existing data-driven battery SOH prediction methods typically use specific health features and different models for fitting or modeling. Health features usually include cycle count, temperature, and charge / discharge capacity. A drawback of these models is that when an individual cell exhibits abnormal degradation, the correlation between various health features and the actual health condition changes, leading to inaccurate model predictions. Summary of the Invention

[0005] This invention provides a method and device for monitoring vehicle batteries, which can accurately determine abnormal cell conditions and abnormal degradation trends, thereby detecting problematic cells earlier and reducing safety risks caused by cell abnormalities and decreased consistency.

[0006] Therefore, the present invention provides the following technical solution:

[0007] A method for monitoring an on-board battery, the method comprising:

[0008] Based on the data of single-cell battery cycling over a certain period of time, a first prediction model and a second prediction model are established respectively.

[0009] Real-time monitoring of vehicle battery data, and prediction based on the first prediction model and the second prediction model respectively, to obtain the first set of SOH values ​​and the second set of SOH values;

[0010] A first data sequence of cumulative charge amount - predicted SOH value is generated based on the first set of SOH values, and a second data sequence of cumulative charge amount - predicted SOH value is generated based on the second set of SOH values.

[0011] Based on the first data sequence and the second data sequence, determine whether the battery cell is abnormal and / or whether the battery cell degradation trend is abnormal.

[0012] Optionally, establishing the first prediction model and the second prediction model respectively includes:

[0013] Collect data on the cyclic use of individual batteries over a certain period of time, and obtain various characteristic parameters based on the cyclic use data;

[0014] A first prediction model and a second prediction model are established based on different feature parameters.

[0015] Optionally, the various characteristic parameters include any one or more of the following: constant current charging time in different voltage ranges, charging current, temperature, duration ratio of temperature ranges, cumulative charging amount, and corresponding battery SOH value;

[0016] The establishment of the first prediction model and the second prediction model based on different feature parameters includes:

[0017] A first prediction model is established based on the cumulative charging amount and the ratio of the duration of the temperature range.

[0018] A second prediction model is established based on the constant current charging time, charging current, and temperature characteristics of the different voltage ranges.

[0019] Optionally, both the first prediction model and the second prediction model are neural network models.

[0020] Optionally, determining whether a battery cell is abnormal based on the first data sequence and the second data sequence includes:

[0021] Calculate the average value of the first data sequence and the average value of the second data sequence from the first cycle to the current cycle.

[0022] Calculate the difference between two means;

[0023] If the difference is greater than a set threshold, it is determined that the battery cell is abnormal.

[0024] Optionally, determining whether the cell degradation trend is abnormal based on the first data sequence and the second data sequence includes:

[0025] Calculate the cosine similarity between the first data sequence and the second data sequence;

[0026] Calculate the similarity index based on the cosine similarity;

[0027] If the similarity index is less than the set similarity threshold, then the cell degradation trend is determined to be abnormal.

[0028] An on-board battery monitoring device, the device comprising: a model building module, a monitoring module, a sequence generation module, and a first judgment module and / or a second judgment module;

[0029] The model building module is used to pre-build a first prediction model and a second prediction model based on the single battery cycle usage data within a certain time period.

[0030] The monitoring module is used to monitor the vehicle battery data in real time and make predictions based on the first prediction model and the second prediction model respectively to obtain a first set of SOH values ​​and a second set of SOH values.

[0031] The sequence generation module is used to generate a first data sequence of cumulative charge amount - predicted SOH value based on the first group of SOH values, and to generate a second data sequence of cumulative charge amount - predicted SOH value based on the second group of SOH values.

[0032] The first judgment module is used to determine whether the battery cell is abnormal based on the first data sequence and the second data sequence;

[0033] The second judgment module is used to determine whether the cell degradation trend is abnormal based on the first data sequence and the second data sequence.

[0034] Optionally, the model building module includes:

[0035] The data acquisition unit is used to collect data on the cyclic use of individual batteries within a certain time period, and to obtain various characteristic parameters based on the cyclic use data;

[0036] The model training unit is used to build a first prediction model and a second prediction model based on different feature parameters, respectively.

[0037] Optionally, the first determination module includes:

[0038] The first calculation unit is used to calculate the average value of the first data sequence and the average value of the second data sequence from the first cycle to the current cycle; and to calculate the difference between the two average values.

[0039] The first judgment unit is used to determine that there is an abnormality in the battery cell when the difference is greater than a set threshold.

[0040] Optionally, the second determination module includes:

[0041] The second calculation unit is used to calculate the cosine similarity between the first data sequence and the second data sequence, and to calculate a similarity index based on the cosine similarity.

[0042] The second judgment unit is used to determine that the cell degradation trend is abnormal when the similarity index is less than the set similarity threshold.

[0043] The vehicle battery monitoring method and device provided by this invention establishes a first prediction model and a second prediction model based on different characteristics of single-cell battery cycle usage data within a certain time period. It monitors vehicle battery data in real time and makes predictions based on the first and second prediction models respectively, obtaining a first set of State of Health (SOH) values ​​and a second set of SOH values. A first data sequence of cumulative charge amount minus predicted SOH value is generated based on the first set of SOH values, and a second data sequence of cumulative charge amount minus predicted SOH value is generated based on the second set of SOH values. The first and second data sequences are used to determine whether the battery cell is abnormal and / or whether the cell degradation trend is abnormal. The vehicle battery monitoring method and device provided by this invention, through two different health characteristic prediction models for related numerical calculations, can accurately screen out battery cells with abnormal degradation trends, identifying problematic battery cells before the battery health status drops to a dangerous range, and reducing safety risks caused by battery cell abnormalities and decreased consistency. Attached Figure Description

[0044] Figure 1 This is a flowchart of an on-board battery monitoring method provided in an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of battery health information extracted through online monitoring in an embodiment of the present invention.

[0046] Figure 3 This is a schematic diagram of a vehicle battery monitoring device provided in an embodiment of the present invention. Detailed Implementation

[0047] To enable those skilled in the art to better understand the embodiments of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and implementation methods.

[0048] The State of Harm (SOH) of a battery is usually calculated as the percentage of the battery's discharge capacity relative to its initial capacity, i.e.:

[0049] SOH=C / C0

[0050] Where C is the discharge capacity of the current battery under standard current, and C0 is the discharge capacity of a brand new battery under standard current.

[0051] Existing battery SOH prediction models typically determine battery health and cell abnormalities by solely predicting the SOH value. This approach cannot identify cells with abnormal degradation trends. To address the shortcomings of existing data-driven battery SOH prediction methods, this invention provides an on-board battery monitoring method and device. By performing relevant numerical calculations using two different health characteristic prediction models, cells with abnormal degradation trends can be screened out, allowing for the early detection of problematic cells before the battery health deteriorates to a dangerous level.

[0052] like Figure 1 The diagram shown is a flowchart of an on-board battery monitoring method provided by an embodiment of the present invention, which includes the following steps:

[0053] Step 101: Based on the data of single-cell battery cycling over a certain period of time, establish a first prediction model and a second prediction model respectively.

[0054] Collect single-cell battery cycle usage data over a certain period and obtain various characteristic parameters based on the cycle usage data. These characteristic parameters may include, but are not limited to, any one or more of the following: constant current charging time in different voltage ranges, charging current, temperature, the proportion of temperature range duration, cumulative charging amount, and the corresponding battery SOH value. Then, establish a first prediction model and a second prediction model based on different characteristic parameters. For example, preprocess the cycle data of a single-cell battery over two years to filter constant current charging time in different voltage ranges, charging current, temperature, the proportion of temperature range duration, cumulative charging amount, and the corresponding battery health status (SOH).

[0055] In one non-limiting embodiment, a first prediction model can be established based on the cumulative charge amount and the proportion of the temperature range duration. The inputs to the first prediction model are the cumulative charge amount and the proportion of the temperature range duration, and the output is the corresponding battery SOH value.

[0056] The Pearson correlation coefficient between the constant current charging time and the corresponding State of Health (SOH) in different voltage ranges of each cycle is calculated. The charging time within the voltage range with the larger correlation coefficient is taken as the health characteristic. Based on the constant current charging time, charging current, and temperature characteristics of different voltage ranges, a second prediction model is established. The inputs of the second prediction model are the constant current charging time, charging current, and temperature in different voltage ranges, and the output is the corresponding battery SOH value.

[0057] In practical applications, the first prediction model and the second prediction model can be neural network models, such as RBF (radial basis function) neural networks, or other types of neural networks and various regression models, etc., which are data-driven prediction models. This embodiment of the invention does not limit the specific types of prediction models.

[0058] Step 102: Monitor the vehicle battery data in real time, and make predictions based on the first prediction model and the second prediction model respectively to obtain the first set of SOH values ​​and the second set of SOH values.

[0059] The vehicle battery data includes, but is not limited to, any one or more of the following: constant current charging time in different voltage ranges, charging current, temperature, duration ratio of temperature ranges, and cumulative charging amount.

[0060] Based on the vehicle battery data obtained from monitoring, predictions were made using the first prediction model and the second prediction model to obtain the first set of SOH values ​​and the second set of SOH values.

[0061] Figure 2 This is a schematic diagram of battery health information extracted through online monitoring in an embodiment of the present invention.

[0062] like Figure 2 As shown, the main data directly monitored by the automotive online platform includes: cell code, data time, vehicle status, charging status, operating status, total mileage, cumulative mileage, total voltage, total current, charge capacity, insulation resistance, accelerator pedal status, brake pedal status, cell temperature, cell voltage, and cell current.

[0063] The main features extracted include: the product of current and time during charging to obtain the charging amount for each time; summing the recorded charging amounts to obtain the cumulative charging amount; recording the proportion of battery temperature range holding time to the total time; when charging is started and the charging current is constant, monitoring the cell voltage value every 1 second and calculating the interval voltage charging time (the optimal voltage range can be obtained by calculating the correlation between different voltage ranges and SOH, generally 3.85V to 4.15V); extracting the current magnitude, average temperature, and calculating the average voltage during this process.

[0064] Step 103: Generate a first data sequence of cumulative charge amount - predicted SOH value based on the first set of SOH values, and generate a second data sequence of cumulative charge amount - predicted SOH value based on the second set of SOH values.

[0065] Specifically, the predicted data is arranged in the order of cumulative charging amount - predicted SOH value to form the prediction result sequence, thereby obtaining the first data sequence L1 corresponding to the first group of SOH values ​​and the second data sequence L2 corresponding to the second group of SOH values.

[0066] Step 104: Determine whether the battery cell is abnormal and / or whether the battery cell degradation trend is abnormal based on the first data sequence and the second data sequence.

[0067] The following methods can be used to determine whether a battery cell is abnormal:

[0068] First, calculate the average value of the first data sequence and the average value of the second data sequence from the first cycle to the current cycle; calculate the difference between the two average values; then determine whether the battery cell is abnormal based on this difference. Specifically, if the difference is greater than a set threshold (e.g., 0.015), the battery cell is determined to be abnormal.

[0069] The average value can be calculated using the following formula:

[0070]

[0071] Where meanSOH is the average health status value, C is the cumulative charge amount, and SOH0 is the battery health status under the current charge capacity.

[0072] The following methods can be used to determine whether the cell degradation trend is abnormal:

[0073] First, the cosine similarity between the first data sequence and the second data sequence is calculated. A similarity index is then calculated based on this cosine similarity, and a judgment is made based on this index. Specifically, if the similarity index is less than a set similarity threshold (e.g., 0.00015), it indicates a significant difference in the curve trends corresponding to the first and second data sequences, thus confirming an abnormal cell degradation trend.

[0074] The formula for calculating the cosine similarity is as follows:

[0075]

[0076] The similarity index is 1-cos(θ).

[0077] Figure 2 A comparison chart of cosine similarity for different decay trends is shown.

[0078] It should be noted that the battery cell mentioned in the embodiments of the present invention refers to a single battery cell in a battery pack.

[0079] The vehicle battery monitoring method provided by this invention establishes a first prediction model and a second prediction model based on different characteristics of single-cell battery cycle usage data within a certain time period. It monitors vehicle battery data in real time and makes predictions based on the first and second prediction models respectively, obtaining a first set of State of Health (SOH) values ​​and a second set of SOH values. A first data sequence of cumulative charge amount minus predicted SOH value is generated based on the first set of SOH values, and a second data sequence of cumulative charge amount minus predicted SOH value is generated based on the second set of SOH values. The first and second data sequences are used to determine whether the battery cell is abnormal and / or whether the cell degradation trend is abnormal. The vehicle battery monitoring method and device provided by this invention perform relevant numerical calculations using two different health characteristic prediction models, thereby accurately screening out battery cells with abnormal degradation trends. This allows for early detection of problematic battery cells before the battery health status drops to a dangerous range, reducing safety risks caused by cell abnormalities and decreased consistency.

[0080] The following examples further illustrate the solution of the present invention in detail.

[0081] Example 1: SOH analysis and anomaly judgment of battery cells during experimental testing

[0082] Step 1: Acquisition of raw training data for test cells

[0083] Multiple long-term cycle experiments were designed for a specific battery model to obtain constant current charging data under different environmental conditions and different battery health statuses. Various data types and corresponding SOH values ​​were extracted from the data.

[0084] Step 2: Training the neural network model for different health characteristics

[0085] The health features in the battery cycle data are classified. The first health feature includes: the cumulative charge amount of the battery and the percentage of time the battery is charged at different temperatures. The second health feature includes: the voltage-constant current charging time during constant current charging of the cell, the charging current magnitude, the average charging temperature during the charging interval, and the average voltage. An RBF radial basis function neural network is used to predict the battery's state of harm (SOH). Two prediction models are trained using health feature 1 and health feature 2 as inputs, respectively, to obtain the first prediction model and the second prediction model.

[0086] Step 3: Online SOH Prediction

[0087] The following parameters were extracted from the battery cells used in the experiment: voltage charging time during the constant current charging process, average charging temperature during the charging period, magnitude of constant current charging current, cumulative battery charge, and the percentage of time the battery spent in different temperature ranges (10-20℃, 20-30℃, 30-40℃). The extracted first and second health features were input into the first and second prediction models, respectively, to obtain the corresponding predicted values ​​SOH1 and SOH2.

[0088] Step 4: Identifying Abnormal Cells

[0089] In battery cycle testing, multiple sets of SOH values ​​are cumulatively predicted for batteries at different cycle periods, forming SOH prediction arrays L1 and L2 for different usage periods. The data points in L1 and L2 represent the prediction curves (or arrays) of different prediction models under the same cumulative charging capacity.

[0090] Calculate the average values ​​of the L1 and L2 sequences respectively. The difference between the two values ​​is calculated, i.e., meanSOH1 - meanSOH2. If the difference between the two mean values ​​is greater than δ, the cell is marked as abnormal, which is used to determine whether the cell capacity is abnormal. The parameter δ is usually set to 0.015.

[0091] The cosine similarity between the two curves is used to determine whether the battery degradation trend is normal. The cosine similarity cos(θ) can be calculated using the formula mentioned earlier. When 1-cos(θ)>0.00015, it can be determined that there is a significant difference between the battery degradation trends of the two prediction models, indicating that the degradation trend of this cell is abnormal.

[0092] Assuming that the prediction results of the two models for the battery cells used in this screening process are δ<0.015 and the similarity index 1-cos(θ)<0.00015, it is determined that the battery cells used in the experiment do not have SOH abnormalities or abnormal capacity decay trends.

[0093] Example 2: On-board battery cell SOH prediction and abnormal cell screening

[0094] Step 1: Acquisition of raw training data

[0095] The training model uses cyclic data from experimentally aged battery cells, the same as in Example 1.

[0096] Step 2: Training the neural network model for different health characteristics

[0097] Different health features are extracted and classified from battery cell experimental data, and then trained to obtain a first prediction model and a second prediction model.

[0098] Step 3: Extracting vehicle battery cell data

[0099] By configuring the vehicle battery monitoring software, the monitoring system samples every 10 seconds while the vehicle is running, sampling data such as individual battery voltage, battery discharge current, data time, total voltage, total mileage, insulation resistance, and individual battery temperature. During charging, sampling occurs every 10 seconds, and within a set constant current charging voltage range, sampling occurs every 5 seconds, sampling data such as individual battery voltage, battery charging current, data time, individual battery temperature, and total charge. When the vehicle is off, sampling occurs every 120 seconds, sampling data such as battery voltage, insulation resistance, and individual battery temperature.

[0100] Step 4: Online SOH Prediction

[0101] Based on the collected data of the vehicle battery cells, health characteristics such as cumulative charging amount, storage ratio in different temperature ranges, constant current charging current, interval voltage charging time, and average charging temperature are selected and calculated. The health characteristics are classified and the SOH value is predicted by different models (i.e., the first prediction model and the second prediction model mentioned above) to obtain the predicted values ​​SOH1 and SOH2.

[0102] Step 5: Identification of Abnormal Battery Cells

[0103] During the use of electric vehicles, multiple sets of SOH1 and SOH2 values ​​are cumulatively predicted for the same battery cell. A first data sequence L1 of cumulative charging amount - predicted SOH value is generated based on the first set of SOH values, and a second data sequence L2 of cumulative charging amount - predicted SOH value is generated based on the second set of SOH values.

[0104] The data points of the first data sequence L1 and the second data sequence L2 represent the prediction curves (or arrays) of different prediction models under the same cumulative charging capacity.

[0105] The vehicle battery module is composed of multiple cells connected in series and parallel. Steps 1-4 are performed on each cell to calculate the L1 and L2 array sets of each cell. The average value and similarity index of L1n and L2n (n is the cell number) of each cell are obtained, which can determine the abnormal condition of each cell in the vehicle module.

[0106] Accordingly, the present invention also provides an on-board battery monitoring device, such as... Figure 3 The diagram shown is a structural schematic of the device.

[0107] In this embodiment, the device 300 includes: a model building module 301, a monitoring module 302, a sequence generation module 303, and a first judgment module 304 and / or a second judgment module 305. Wherein:

[0108] The model building module 301 is used to pre-build a first prediction model and a second prediction model based on the single battery cycle usage data within a certain time period.

[0109] The monitoring module 302 is used to monitor the vehicle battery data in real time and make predictions based on the first prediction model and the second prediction model respectively to obtain a first set of SOH values ​​and a second set of SOH values.

[0110] The sequence generation module 303 is used to generate a first data sequence of cumulative charge amount - predicted SOH value based on the first group of SOH values, and to generate a second data sequence of cumulative charge amount - predicted SOH value based on the second group of SOH values.

[0111] The first judgment module 304 is used to determine whether the battery cell is abnormal based on the first data sequence and the second data sequence;

[0112] The second judgment module 305 is used to determine whether the cell degradation trend is abnormal based on the first data sequence and the second data sequence.

[0113] In specific applications, a non-limiting structure of the model building module 301 may include the following units:

[0114] The data acquisition unit is used to collect data on the cyclic use of individual batteries within a certain time period, and to obtain various characteristic parameters based on the cyclic use data;

[0115] The model training unit is used to build a first prediction model and a second prediction model based on different feature parameters, respectively.

[0116] Specifically, the first determination module may include the following units:

[0117] The first calculation unit is used to calculate the average value of the first data sequence and the average value of the second data sequence from the first cycle to the current cycle; and to calculate the difference between the two average values.

[0118] The first judgment unit is used to determine that there is an abnormality in the battery cell when the difference is greater than a set threshold.

[0119] Specifically, the second determination module may include the following units:

[0120] The second calculation unit is used to calculate the cosine similarity index between the first data sequence and the second data sequence;

[0121] The second judgment unit is used to determine that the cell degradation trend is abnormal when the cosine similarity index is less than the set similarity threshold.

[0122] The specific implementation methods of the above modules and units can be referred to the description in the previous embodiments of the present invention, and will not be repeated here.

[0123] The vehicle battery monitoring device provided by this invention establishes a first prediction model and a second prediction model based on different characteristics of single-cell battery cycle usage data over a certain period of time. It monitors vehicle battery data in real time and makes predictions based on the first and second prediction models to obtain a first set of State of Health (SOH) values ​​and a second set of SOH values. A first data sequence of cumulative charge amount minus predicted SOH value is generated based on the first set of SOH values, and a second data sequence of cumulative charge amount minus predicted SOH value is generated based on the second set of SOH values. The first and second data sequences are used to determine whether the battery cell is abnormal and / or whether the cell degradation trend is abnormal. The vehicle battery monitoring method and device provided by this invention perform relevant numerical calculations using two different health characteristic prediction models, thereby accurately screening out battery cells with abnormal degradation trends. This allows for the early detection of problematic battery cells before the battery health condition drops to a dangerous range, reducing safety risks caused by cell abnormalities and decreased consistency.

[0124] It should be noted that the terms "comprising" and "having" and any variations thereof in the specification, claims and accompanying drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products or devices.

[0125] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. Furthermore, the system embodiments described above are merely illustrative. The modules and units described as separate components may or may not be physically separate; that is, they may be located on a single network unit or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0126] The embodiments of the present invention have been described in detail above. Specific implementation methods have been used to illustrate the present invention. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and systems of the present invention, and are merely some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention, and the content of this specification should not be construed as a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for monitoring vehicle batteries, characterized in that, The method includes: Based on the data of single-cell battery cycling over a certain period of time, a first prediction model and a second prediction model are established respectively. Real-time monitoring of vehicle battery data, and prediction based on the first prediction model and the second prediction model respectively, to obtain the first set of SOH values ​​and the second set of SOH values; A first data sequence of cumulative charge amount - predicted SOH value is generated based on the first set of SOH values, and a second data sequence of cumulative charge amount - predicted SOH value is generated based on the second set of SOH values. Determine whether the battery cell is abnormal and / or whether the battery cell degradation trend is abnormal based on the first data sequence and the second data sequence; The step of determining whether the battery cell is abnormal based on the first data sequence and the second data sequence includes: Calculate the average value of the first data sequence and the average value of the second data sequence from the first cycle to the current cycle. Calculate the difference between two means; If the difference is greater than a set threshold, it is determined that the battery cell is abnormal; The establishment of the first prediction model and the second prediction model respectively includes: Collect data on the cyclic use of individual batteries over a certain period of time, and obtain various characteristic parameters based on the cyclic use data; A first prediction model and a second prediction model are established based on different feature parameters, respectively. The various characteristic parameters include any one or more of the following: constant current charging time in different voltage ranges, charging current, temperature, duration ratio of temperature ranges, cumulative charging amount, and corresponding battery SOH value; The establishment of the first prediction model and the second prediction model based on different feature parameters includes: A first prediction model is established based on the cumulative charging amount and the ratio of the duration of the temperature range. A second prediction model is established based on the constant current charging time, charging current, and temperature characteristics of the different voltage ranges.

2. The method according to claim 1, characterized in that, Both the first and second prediction models are neural network models.

3. The method according to any one of claims 1 to 2, characterized in that, The step of determining whether the cell degradation trend is abnormal based on the first data sequence and the second data sequence includes: Calculate the cosine similarity between the first data sequence and the second data sequence; Calculate the similarity index based on the cosine similarity; If the similarity index is less than the set similarity threshold, then the cell degradation trend is determined to be abnormal.

4. An apparatus for use in the vehicle battery monitoring method according to any one of claims 1-3, characterized in that, The device includes: a model building module, a monitoring module, a sequence generation module, and a first judgment module and / or a second judgment module; The model building module is used to pre-build a first prediction model and a second prediction model based on the single battery cycle usage data within a certain time period. The monitoring module is used to monitor the vehicle battery data in real time and make predictions based on the first prediction model and the second prediction model respectively to obtain a first set of SOH values ​​and a second set of SOH values. The sequence generation module is used to generate a first data sequence of cumulative charge amount - predicted SOH value based on the first group of SOH values, and to generate a second data sequence of cumulative charge amount - predicted SOH value based on the second group of SOH values. The first judgment module is used to determine whether the battery cell is abnormal based on the first data sequence and the second data sequence; The second judgment module is used to determine whether the cell degradation trend is abnormal based on the first data sequence and the second data sequence; The establishment of the first prediction model and the second prediction model respectively includes: Collect data on the cyclic use of individual batteries over a certain period of time, and obtain various characteristic parameters based on the cyclic use data; A first prediction model and a second prediction model are established based on different feature parameters, respectively. The various characteristic parameters include any one or more of the following: constant current charging time in different voltage ranges, charging current, temperature, duration ratio of temperature ranges, cumulative charging amount, and corresponding battery SOH value; The establishment of the first prediction model and the second prediction model based on different feature parameters includes: A first prediction model is established based on the cumulative charging amount and the ratio of the duration of the temperature range. A second prediction model is established based on the constant current charging time, charging current, and temperature characteristics of the different voltage ranges.

5. The apparatus according to claim 4, characterized in that, The model building module includes: The data acquisition unit is used to collect data on the cyclic use of individual batteries within a certain time period, and to obtain various characteristic parameters based on the cyclic use data; The model training unit is used to build a first prediction model and a second prediction model based on different feature parameters, respectively.

6. The apparatus according to claim 4 or 5, characterized in that, The first judgment module includes: The first calculation unit is used to calculate the average value of the first data sequence and the average value of the second data sequence from the first cycle to the current cycle; and to calculate the difference between the two average values. The first judgment unit is used to determine that there is an abnormality in the battery cell when the difference is greater than a set threshold.

7. The apparatus according to claim 4 or 5, characterized in that, The second judgment module includes: The second calculation unit is used to calculate the cosine similarity between the first data sequence and the second data sequence, and to calculate a similarity index based on the cosine similarity. The second judgment unit is used to determine that the cell degradation trend is abnormal when the similarity index is less than the set similarity threshold.

Images