A method and apparatus for identifying battery model parameters

Abstract

This invention provides a method and apparatus for identifying battery model parameters. It reads battery data within a specific time period from a cloud platform database, extracts multiple charge / discharge segments from each data point, and performs a Discrete Fourier Transform on each segment in descending order of timestamps to generate a spectrum. The frequency range of each charge / discharge segment is determined based on the spectrum, and filtering is performed to obtain the target charge / discharge segment data. Charging and discharging parameters are estimated based on the target charge / discharge segment data and the preset frequency range. This solution does not rely on large-scale training sets or long-term training, and can obtain battery model parameters with high accuracy without the need for neural network algorithms, ensuring real-time parameter updates from the cloud-based BMS to the vehicle-side BMS. Frequency analysis can reveal parameter changes throughout the battery's entire lifespan, meeting the needs of fault diagnosis and safety warnings over long time scales, and enriching the battery parameter identification methods for cloud-based BMS.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence and big data technology, and in particular to a method and apparatus for identifying battery model parameters. Background Technology

[0002] Battery Management Systems (BMS), as the core of battery systems, are widely used in various new energy vehicles and energy storage power stations. BMS monitors the state of the battery system and ensures its safe operation by establishing battery models and calculating model parameters. With the introduction of big data technology, the development of BMS has gradually entered the cloud stage, bringing innovation to battery model parameter identification algorithms.

[0003] Currently, most cloud-based BMS application layer algorithms for new energy vehicles employ pre-trained algorithms such as neural networks and machine learning. The accuracy of these algorithms is strongly correlated with the size and quality of the training dataset, which limits their coverage to some extent. Secondly, these algorithms cover more aspects of fault diagnosis, safety warnings, and state estimation, but less about parameter identification. Furthermore, the development of artificial intelligence and big data will inevitably lead to more and more functions being added to cloud-based BMS for new energy vehicles. Large-scale neural network training will consume a lot of cloud computing resources, which will have a certain impact on the real-time performance of the cloud-based BMS.

[0004] Therefore, improving the efficiency and real-time performance of battery model parameter identification algorithms in cloud-based BMS has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a method and apparatus for identifying battery model parameters, in order to solve the problems of high dependence on training data and poor real-time performance in current battery model parameter identification.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of this invention discloses a method for identifying battery model parameters, the method comprising:

[0008] Read multiple battery data entries within a preset time period from the cloud platform database;

[0009] If multiple battery data sets meet the identification requirements, then multiple charge / discharge segment data sets are determined based on the voltage, current sign, charging flag status, and timestamp in each battery data set; each charge / discharge segment data set corresponds to a target timestamp.

[0010] According to the target timestamps in descending order, perform discrete Fourier transform on all charge and discharge segment data respectively to obtain the spectrum diagram corresponding to each charge and discharge segment data;

[0011] Based on the spectrum diagram, the preset frequency threshold range to which each charge / discharge segment data belongs is determined, and all charge / discharge segment data is filtered using a filter to obtain multiple target charge / discharge segment data.

[0012] If any of the target charge / discharge segment data belongs to a preset frequency threshold range that is a low frequency range, the target charge / discharge segment data is marked as the first charge / discharge segment data, and the battery ohmic internal resistance is calculated based on the first charge / discharge segment data and the state equation of the battery nth order RC model.

[0013] Based on the battery's ohmic internal resistance and all second charge / discharge segment data, multiple charging parameters and multiple discharging parameters are estimated; the preset frequency threshold range to which the second charge / discharge segment data belongs is a non-low frequency range, and the target timestamp corresponding to the second charge / discharge segment data is greater than the target timestamp corresponding to the first charge / discharge segment data.

[0014] Preferably, if multiple battery data sets meet the identification requirements, before determining multiple charge / discharge segment data based on the voltage, current sign, charging flag status, and timestamp in each battery data set, the method further includes:

[0015] Check if the first battery data exists in all battery data, where the charging flag is on.

[0016] If the first battery data exists, then check if the second battery data exists among all battery data; the timestamp carried by the second battery data is greater than the timestamp carried by the first battery data, and the charging flag in the second battery data is in the off state;

[0017] If a second set of battery data exists, then multiple sets of battery data are determined to meet the identification requirements.

[0018] Preferably, the step of determining multiple charge / discharge segment data based on the voltage, current sign, charging flag status, and timestamp in each piece of battery data includes:

[0019] Calculate the direction of voltage change gradient based on the voltage and timestamp in each piece of battery data;

[0020] Multiple charge / discharge segment data are determined by combining the current sign, charging flag status, and voltage change gradient direction of each battery data entry; each charge / discharge segment data includes multiple battery data entries.

[0021] For each charge / discharge segment data, the timestamp with the smallest value among the multiple battery data contained in the charge / discharge segment data is marked as the target timestamp corresponding to the charge / discharge segment data.

[0022] Preferably, determining the preset frequency threshold range to which each charge / discharge segment data belongs based on the spectrum includes:

[0023] Obtain multiple preset frequency threshold ranges;

[0024] For each charge / discharge segment data, a spectrum diagram of the charge / discharge segment data is obtained; the spectrum diagram includes multiple frequency components.

[0025] For each preset frequency threshold interval, the number of frequency components located in the preset frequency threshold interval in the spectrum is calculated, and the ratio between the number of frequency components and the total number of frequency components in the spectrum is calculated.

[0026] If the ratio is greater than a preset percentage threshold, then the charge / discharge segment data is determined to belong to the preset frequency threshold range, and the preset frequency threshold range to which each charge / discharge segment data belongs is obtained.

[0027] Preferably, the step of calculating the battery's ohmic internal resistance based on the first charge / discharge segment data and the battery's nth-order RC model state equation includes:

[0028] Obtain the state equation of the nth-order RC model of the battery;

[0029] By setting the value of n to 1, the state equation of the nth-order RC model of the battery is simplified to obtain the state equation of the pure resistance model.

[0030] Based on the state equation of the pure resistance model and the data from the first charge / discharge segment, the ohmic internal resistance of the battery is calculated.

[0031] A second aspect of this invention discloses a battery model parameter identification device, the device comprising:

[0032] The reading unit is used to read multiple battery data entries within a preset time period from the cloud platform database.

[0033] The first determining unit is configured to, if multiple pieces of battery data meet the identification requirements, determine multiple charge / discharge segment data based on the voltage, current sign, charging flag status, and timestamp in each piece of battery data; each charge / discharge segment data corresponds to a target timestamp.

[0034] The Discrete Fourier Transform (DFT) processing unit is used to perform DFT processing on all charge-discharge segment data in descending order of the target timestamps to obtain the spectrum diagram corresponding to each charge-discharge segment data.

[0035] The filtering processing unit is used to determine the preset frequency threshold range to which each charge-discharge segment data belongs based on the spectrum diagram, and to use a filter to filter all charge-discharge segment data to obtain multiple target charge-discharge segment data.

[0036] The calculation unit is used to mark the target charge-discharge segment data as the first charge-discharge segment data if the preset frequency threshold interval to which any of the target charge-discharge segment data belongs is a low frequency interval, and to calculate the battery ohmic internal resistance based on the first charge-discharge segment data and the state equation of the battery nth order RC model.

[0037] The estimation unit is used to estimate multiple charging parameters and multiple discharging parameters based on the battery's ohmic internal resistance and all second charge-discharge segment data; the preset frequency threshold range to which the second charge-discharge segment data belongs is a non-low frequency range, and the target timestamp corresponding to the second charge-discharge segment data is greater than the target timestamp corresponding to the first charge-discharge segment data.

[0038] Preferably, the device further includes:

[0039] The first query unit is used to query whether there is a first battery data in all battery data, where the charging flag in the first battery data is in the on state.

[0040] The second query unit is used to query whether second battery data exists among all battery data if first battery data exists; the timestamp carried by the second battery data is greater than the timestamp carried by the first battery data, and the charging flag in the second battery data is in the off state;

[0041] The second determining unit is used to determine that multiple sets of battery data satisfy the identification requirements if second battery data exists.

[0042] Preferably, the first determining unit includes:

[0043] The calculation module is used to calculate the direction of voltage change gradient based on the voltage and timestamp in each piece of battery data;

[0044] The determination module is used to determine multiple charge / discharge segment data by combining the current sign, charging flag status, and voltage change gradient direction of each piece of battery data; each piece of charge / discharge segment data includes multiple pieces of battery data;

[0045] The marking module is used to mark the timestamp with the smallest value among the multiple battery data contained in each charge / discharge segment data as the target timestamp corresponding to that charge / discharge segment data.

[0046] Preferably, the filtering unit is specifically used for:

[0047] Multiple preset frequency threshold intervals are obtained; for each charge / discharge segment data, a spectrum of the charge / discharge segment data is obtained; the spectrum includes multiple frequency components; for each preset frequency threshold interval, the number of frequency components located in the preset frequency threshold interval in the spectrum is calculated, and the ratio between the number of frequency components and the total number of frequency components in the spectrum is calculated.

[0048] If the ratio is greater than a preset percentage threshold, then the charge / discharge segment data is determined to belong to the preset frequency threshold range, and the preset frequency threshold range to which each charge / discharge segment data belongs is obtained.

[0049] Preferably, the computing unit is specifically used for:

[0050] Obtain the state equation of the nth-order RC model of the battery; set the value of n to 1, simplify the state equation of the nth-order RC model of the battery to obtain the state equation of the pure resistance model; calculate the ohmic internal resistance of the battery based on the state equation of the pure resistance model and the first charge and discharge segment data.

[0051] Based on the battery model parameter identification method and apparatus provided in the above embodiments of the present invention, battery data within a specific time period is read from a cloud platform database. Multiple charge / discharge segment data are extracted from each battery data entry. The data is then sorted in descending order of timestamp and subjected to a Discrete Fourier Transform to generate a spectrum. The frequency range to which each charge / discharge segment data belongs is determined based on the spectrum, and filtering is performed to obtain the target charge / discharge segment data. Charging and discharging parameters are estimated based on the target charge / discharge segment data and the preset frequency range. This solution does not rely on large-scale training sets or long-term training, and can obtain battery model parameters with high accuracy without the need for neural network algorithms, ensuring real-time parameter updates from the cloud-based BMS to the vehicle-side BMS. Frequency analysis can reveal parameter changes throughout the battery's entire lifespan, meeting the needs of fault diagnosis and safety warning over long time scales, and enriching the battery parameter identification methods for cloud-based BMS. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0053] Figure 1 A flowchart illustrating a method for identifying battery model parameters provided in an embodiment of the present invention;

[0054] Figure 2 A flowchart for determining multiple charge / discharge segment data provided in an embodiment of the present invention;

[0055] Figure 3 A first schematic diagram of a battery model parameter identification method provided in an embodiment of the present invention;

[0056] Figure 4 This is a schematic diagram illustrating the conversion of various charge / discharge segment data from the time domain to the frequency domain, as provided in the embodiments of the present invention.

[0057] Figure 5 A schematic diagram of a method for determining the preset frequency threshold range to which each charge / discharge segment data belongs, provided in an embodiment of the present invention;

[0058] Figure 6 This is a flowchart provided by an embodiment of the present invention for determining the preset frequency threshold range to which each charge / discharge segment data belongs based on a spectrum diagram;

[0059] Figure 7 This is a schematic diagram illustrating the process of calculating the ohmic internal resistance of a battery, provided in an embodiment of the present invention.

[0060] Figure 8 A schematic diagram of an nth-order RC model of a battery provided in an embodiment of the present invention;

[0061] Figure 9 This is a structural block diagram of a battery model parameter identification device provided in an embodiment of the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] In this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0064] As the background technology indicates, the cloud-based BMS application layer algorithm for new energy vehicles suffers from several issues, including reliance on high-quality training data, a lack of effective battery parameter identification methods, real-time performance problems due to high demands on cloud computing resources, and poor interpretability of deep learning models. These issues may affect the accuracy, response speed, and maintenance efficiency of the battery management system, necessitating improvements through algorithm optimization, enhanced data quality, and improved model interpretability.

[0065] Therefore, this invention provides a method and apparatus for identifying battery model parameters. It reads battery data within a specific time period from a cloud platform database, extracts multiple charge / discharge segment data from each data point, and performs a discrete Fourier transform on each segment data point in descending order of timestamp to generate a spectrum. Based on the spectrum, the frequency range to which each charge / discharge segment data belongs is determined, and filtering is performed to obtain the target charge / discharge segment data. Charging and discharging parameters are estimated based on the target charge / discharge segment data and the preset frequency range. This solution does not rely on large-scale training sets or long-term training, and can obtain battery model parameters with high accuracy without the need for neural network algorithms, reducing cloud computing power requirements and ensuring real-time parameter updates from the cloud-based BMS to the vehicle-side BMS. Furthermore, frequency analysis can reveal parameter changes throughout the battery's entire lifespan, meeting the needs of fault diagnosis and safety warnings over long time scales, and enriching the battery parameter identification methods for cloud-based BMS.

[0066] See Figure 1 The flowchart illustrates a method for identifying battery model parameters according to an embodiment of the present invention. The method includes:

[0067] Step S101: Read multiple battery data entries within a preset time period from the cloud platform database.

[0068] It should be noted that the cloud platform database includes multiple records of real vehicle operation data.

[0069] In the specific implementation step S101, multiple battery data points within a preset time period are read from multiple real vehicle operation data points in the cloud platform database. Each battery data point includes, but is not limited to, voltage data, current data, charging status, and timestamp.

[0070] For example, read multiple battery data points for 24 hours from multiple real vehicle operation data points in the cloud platform database.

[0071] Understandably, collecting 24-hour battery data for parameter identification can cover battery performance under different operating conditions, such as charging and discharging processes and temperature changes. This helps capture changes in battery state over long-term use, improving the accuracy and reliability of the model. Long-term data collection can identify potential battery problems under different operating conditions, such as the impact of temperature fluctuations or prolonged high loads, thereby enhancing the stability of parameter identification results and providing support for battery management system optimization.

[0072] It should be noted that the collected battery data points need to be evaluated to determine whether they meet the identification criteria, ensuring that they effectively reflect the true state of the battery and thus providing a reliable basis for subsequent parameter identification. Based on this, refer to processes A1 to A3 shown in the embodiments of this invention for a detailed explanation of the specific content of evaluating whether multiple battery data points meet the identification criteria.

[0073] Process A1: Query whether the first battery data exists in all battery data.

[0074] It should be noted that the charging indicator in the first battery data is in the "on" state.

[0075] In the specific implementation process A1, check if there is a first battery data in all battery data with the charging flag on; if the first battery data exists, then execute process A2; if the first battery data does not exist, then end the process.

[0076] Process A2: If the first battery data exists, then query all battery data to see if the second battery data exists.

[0077] It should be noted that the timestamp carried by the second battery data is greater than the timestamp carried by the first battery data, and the charging flag in the second battery data is in the off state.

[0078] In the specific implementation of process A2, if the first battery data exists, then check if the second battery data exists among all battery data; if the second battery data exists, then execute process A3; if the second battery data does not exist, then end the process.

[0079] Process A3: If a second battery data exists, then determine that multiple battery data meet the identification requirements.

[0080] In other words, when multiple battery data entries include complete charging process data, it is determined that the multiple battery data entries meet the identification requirements.

[0081] Understandably, complete charging process data can provide full-cycle data of the battery from low charge to full charge, covering different operating states of the battery, thereby improving the accuracy and reliability of parameter identification.

[0082] Step S102: If multiple battery data meet the identification requirements, then determine multiple charge / discharge segment data based on the voltage, current symbol, charging flag status, and timestamp in each battery data.

[0083] In the specific implementation step S102, when multiple battery data entries read from the cloud platform database meet the identification requirements, multiple charge / discharge segment data are determined from all battery data based on the voltage, current symbols, charging flag status, and timestamp in each battery data entry. Specifically, these multiple charge / discharge segment data include multiple charging segment data and multiple discharging segment data.

[0084] It is understood that the specific process for determining multiple charge / discharge segment data is detailed in the embodiments of this invention. Figure 2 The content in Figure 2 include:

[0085] Step S201: Calculate the direction of voltage change gradient based on the voltage and timestamp in each battery data entry.

[0086] In the specific implementation of step S201, the voltage change of all battery data over time is calculated based on the voltage and timestamp in each battery data, and the direction of the voltage change gradient is determined based on the sign of the voltage change.

[0087] Step S202: Combine the current sign, charging flag status, and voltage change gradient direction of each battery data to determine multiple charge / discharge segment data.

[0088] Understandably, each battery data entry defines charging current and discharging current with different signs; for example, the current sign is negative when charging and positive when discharging.

[0089] In the specific implementation step S202, based on the current sign, charging flag status, and voltage change gradient direction of each battery data segment (i.e., as the time represented by the timestamp changes), the system analyzes whether the voltage increases or decreases, whether the current is charging or discharging, and whether the charging flag is in the on state, to determine multiple charging segment data (e.g., B1, B2, ..., B...). i B n ) and multiple discharge segment data (e.g., C1, C2, ..., C i ..., C n ).

[0090] It should be noted that multiple charging segment data and multiple discharging segment data are represented as multiple charge and discharge segment data.

[0091] It is understandable that each charge / discharge segment data includes multiple battery data points, meaning that each charge / discharge segment data represents battery data over a continuous period of time.

[0092] Step S203: For each charge / discharge segment data, mark the timestamp with the smallest value among the multiple battery data contained in the charge / discharge segment data as the target timestamp corresponding to the charge / discharge segment data.

[0093] In the specific implementation of step S203, for each charge / discharge segment data, the timestamp corresponding to the first battery data (i.e., the timestamp with the smallest value) among the multiple battery data contained in the charge / discharge segment data is marked as the target timestamp corresponding to the charge / discharge segment data.

[0094] In other words, each charge / discharge segment of data corresponds to a target timestamp.

[0095] Understandably, identifying multiple charge / discharge segments from multiple battery data points can improve identification accuracy, enable real-time and online identification, reduce computational resource consumption, and improve identification accuracy.

[0096] Step S103: Perform Discrete Fourier Transform on all charge and discharge segment data in descending order of target timestamps to obtain the spectrum diagram corresponding to each charge and discharge segment data.

[0097] In the specific implementation step S103, all charge and discharge segment data are input into the system as described in the embodiment of the present invention, according to the target timestamp in descending order. Figure 3 In the actuator 1 shown, discrete Fourier transform processing is performed to analyze the frequency of each charge-discharge segment data, and the spectrum diagram corresponding to each charge-discharge segment data is obtained.

[0098] It should be noted that, as in the embodiments of the present invention Figure 3 In the actuator 1 shown, the discrete Fourier transform is used to transform the data of each charging and discharging segment, converting it from the time domain to the frequency domain (e.g., in the embodiment of this invention). Figure 4 (As shown).

[0099] For example: All charging segment data (B1, B2, ..., B...) from the charging and discharging segment data. i B n Input the target timestamps in descending order, as shown in the embodiment of the present invention. Figure 3 In the actuator 1 shown, the spectrum diagrams (BSpe1, BSpe2, ..., BSpe) corresponding to each charging segment data are obtained. n ).

[0100] For example: All discharge segment data (C1, C2, ..., C...) from all charge / discharge segment data. i ..., C n Input the target timestamps in descending order, as shown in the embodiment of the present invention. Figure 3 In the actuator 1 shown, the spectrum diagrams (CSpe1, CSpe2, ..., CSpe) corresponding to each charging segment data are obtained. n ).

[0101] In practical applications, other transformation methods can also be used to perform frequency analysis on the data of each charging and discharging segment. The Discrete Fourier Transform proposed in this embodiment of the invention is only for illustrative purposes.

[0102] Step S104: Determine the preset frequency threshold range to which each charge / discharge segment data belongs based on the spectrum diagram, and use a filter to filter all charge / discharge segment data to obtain multiple target charge / discharge segment data.

[0103] Combined with the embodiments of the present invention Figure 5 The content shown is based on the spectrum diagram corresponding to each charge and discharge segment data, multiple preset frequency threshold intervals, and data frequency division criteria. The preset frequency threshold interval to which each charge and discharge segment data belongs is determined, and all charge and discharge segment data are filtered using a filter to obtain multiple target charge and discharge segment data.

[0104] It is understandable that filters (such as high-pass filters, low-pass filters, band-pass filters, band-stop filters, etc.) are used to filter all charge and discharge segment data belonging to each preset frequency threshold range, so that each charge and discharge segment data retains only the frequency components that meet the preset frequency threshold range, while filtering out other frequency components that do not meet the preset frequency threshold range, thus obtaining multiple target charge and discharge segment data.

[0105] In other words, the multiple target charge / discharge segment data includes multiple target discharge segment data Cf1, Cf2, ..., Cf n ; and multiple target charging segment data Bf1, Bf2, ..., Bf n .

[0106] It should be noted that each target charge / discharge segment data corresponds to a preset frequency threshold range to which each charge / discharge segment data belongs.

[0107] It is understood that the specific process of determining the preset frequency threshold interval to which each charge / discharge segment data belongs, based on the spectrum diagram corresponding to each charge / discharge segment data, multiple preset frequency threshold intervals, and data frequency division criteria, is detailed in the embodiments of this invention. Figure 6 The content in Figure 6 include:

[0108] Understandably, data frequency division criteria include setting a preset percentage threshold for the spectrum graph as Ratio, and the spectrum graph (BSpe) i or CSpe i When the ratio of frequency components falling within a certain frequency threshold range to the total number of frequency components in a spectrum is greater than or equal to Ratio, the spectrum is considered valid. i or CSpe i The corresponding charge / discharge segment data B i (or C) i It falls within the frequency threshold range.

[0109] Step S601: Obtain multiple preset frequency threshold ranges.

[0110] It should be noted that multiple preset frequency threshold intervals, such as [0,ω1], (ω1,ω2], (ω2,ω3], ..., (ω...), are used. m (,+∞).

[0111] It should be noted that [0,ω1] is defined as the low-frequency range.

[0112] Step S602: For each charge / discharge segment data, obtain the spectrum of the charge / discharge segment data.

[0113] It is understandable that steps S602 to S605 are performed for each charge / discharge segment data, i.e., each charging segment data and each discharging segment data.

[0114] It should be noted that the spectrum of each charge / discharge segment data includes multiple frequency components.

[0115] Step S603: For each preset frequency threshold interval, calculate the number of frequency components in the spectrum that are located in the preset frequency threshold interval, and calculate the ratio between the number of frequency components and the total number of frequency components in the spectrum.

[0116] It is understandable that for each preset frequency threshold interval [0,ω1], (ω1,ω2], (ω2,ω3], ..., (ω... m ,+∞), calculate the number of frequency components in the spectrum of each charging segment data that are located in the preset frequency threshold range, and calculate the ratio between the number of frequency components and the total number of frequency components in the spectrum of the current charging segment data.

[0117] It should be noted that the ratio is determined to be greater than the preset percentage threshold Ratio; if the ratio is greater than the preset percentage threshold Ratio, then step S604 is executed.

[0118] Step S604: If the ratio is greater than the preset percentage threshold, then the charge / discharge segment data is determined to belong to the preset frequency threshold range, and the preset frequency threshold range to which each charge / discharge segment data belongs is obtained.

[0119] In the specific implementation of step S604, when the ratio is greater than the preset proportion threshold Ratio, it is determined that the current charge / discharge segment data belongs to the current preset frequency threshold range, and finally the preset frequency threshold range to which each charge / discharge segment data belongs is obtained.

[0120] Step S105: If the preset frequency threshold range to which any target charge / discharge segment data belongs is a low frequency range, mark the target charge / discharge segment data as the first charge / discharge segment data, and calculate the battery ohmic internal resistance based on the first charge / discharge segment data and the state equation of the battery nth order RC model.

[0121] It is understandable that [0, ω1] is predefined as the low-frequency range. This is in conjunction with embodiments of the present invention. Figure 3 The content shown demonstrates the use of actuator 2 to identify and classify battery model parameters. The following is a detailed explanation of steps S105 and S106.

[0122] Combined with the embodiments of the present invention Figure 7 The diagram illustrates two methods for calculating the battery's ohmic internal resistance: One method involves marking any target discharge segment data point within a preset frequency threshold range that is a low-frequency range as the first discharge segment data point, and then calculating the battery's ohmic internal resistance based on the first discharge segment data point and the battery's nth-order RC model state equation. The other method involves marking any target charging segment data point within a preset frequency threshold range that is a low-frequency range, and then calculating the battery's ohmic internal resistance based on the first charging segment data point and the battery's nth-order RC model state equation.

[0123] Understandably, the process involves simplifying the state equations of the battery's nth-order RC model and using parameter identification algorithms (including but not limited to Kalman filtering, recursive least squares, and objective optimization algorithms) to calculate the battery's ohmic internal resistance.

[0124] It should be noted that the specific method for simplifying the state equation of the nth-order RC model of the battery and calculating the ohmic internal resistance of the battery is detailed in processes B1 to B3 shown in the embodiments of this invention.

[0125] Process B1: Obtain the state equation of the nth-order RC model of the battery.

[0126] It should be noted that the state equation of the nth-order RC model of the battery (i.e., the state equation of the n-RC model of the battery) is as shown in Equation (1).

[0127] (1)

[0128] Where i represents the index of the RC network (i=1,2,3,…,n), R p C represents the internal polarization resistance of the RC network. p t represents the polarization voltage within the RC network; t represents time. R represents the polarization voltage. p,i C represents the polarization resistance. p,i Let I(t) represent the polarization capacitance, U(t) represent the current, OCV(SOC(t)) represent the open-circuit voltage (a function of SOC), and R0 represent the battery's internal resistance in ohms.

[0129] Process B2: Set the value of n to 1, simplify the state equation of the nth order RC model of the battery, and obtain the state equation of the pure resistance model.

[0130] It is understood that, in conjunction with the embodiments of the present invention Figure 8 As shown, under low-frequency excitation, the capacitor element of the RC network is considered to be approximately short-circuited, and the 1-RC model can be simplified to a pure resistance model. Therefore, the simplified state equation of the battery 1-RC model (i.e., the state equation of the pure resistance model) is as shown in formula (2).

[0131] (2)

[0132] Among them, Q max SOC(t) represents the maximum usable capacity of the battery, SOC(t) represents the SOC value at time t, and SOC(0) represents the SOC value at the start time (i.e., the SOC value at time 0).

[0133] Process B3: Based on the state equation of the pure resistance model and the data of the first charge and discharge segment, the ohmic internal resistance of the battery is calculated.

[0134] In the specific implementation process B3, the battery ohmic internal resistance R0 is calculated based on the pure resistance model state equation shown in formula (2) and the data of the first charge and discharge segment.

[0135] Step S106: Estimate the charging and discharging parameters based on the battery's internal ohmic resistance and all data from the second charge / discharge segments.

[0136] It should be noted that the preset frequency threshold range to which the second charge / discharge segment data belongs is a non-low frequency range, and the target timestamp corresponding to the second charge / discharge segment data is greater than the target timestamp of the first charge / discharge segment data.

[0137] In the specific implementation step S106, the battery ohmic internal resistance R0 is regarded as a known quantity. A parameter identification algorithm (including but not limited to Kalman filtering, recursive least squares, and objective optimization algorithms) is used to estimate the RC network parameters with high responsivity to a certain frequency range in the battery model based on the different RC network responses of each RC network in the n-RC model to each frequency range, combined with all the second charge and discharge segment data, thereby obtaining multiple charging parameters and multiple discharging parameters.

[0138] In some specific embodiments, see Figure 3 and Figure 7 As shown in the embodiments of the present invention, multiple charging parameters and multiple discharging parameters can also be stored in a cloud platform database.

[0139] In this embodiment of the invention, the solution obtains battery charging and discharging data within a specific time period from a cloud platform database through a cloud-based BMS, extracts and arranges each charging and discharging segment in chronological order, and adds a timestamp to each data segment. Then, this data is converted from the time domain to the frequency domain to generate a frequency spectrum, and the data is divided according to a set frequency range. Compared with existing cloud-based BMS algorithms, this solution does not rely on large-scale, high-quality training sets and long-term training, and can obtain battery model parameters with high accuracy without neural network algorithms, reducing the demand for cloud computing power and ensuring real-time parameter updates from the cloud-based BMS to the vehicle-side BMS. Simultaneously, frequency analysis can reveal parameter changes in the battery throughout its entire life cycle, meeting the needs of fault diagnosis and safety warning on a long-term scale, and enriching the battery parameter identification methods of cloud-based BMS.

[0140] Corresponding to the battery model parameter identification method proposed in the above embodiments of the present invention, see [link to relevant documentation]. Figure 9 The diagram illustrates a structural block diagram of a battery model parameter identification device provided in an embodiment of the present invention. The device includes: a reading unit 901, a first determining unit 902, a discrete Fourier transform processing unit 903, a filtering processing unit 904, a calculation unit 905, and an estimation unit 906.

[0141] The reading unit 901 is used to read multiple battery data entries within a preset time period from the cloud platform database.

[0142] The first determining unit 902 is used to determine multiple charge / discharge segment data based on the voltage, current symbol, charging flag status and timestamp in each battery data if multiple battery data meet the identification requirements; each charge / discharge segment data corresponds to a target timestamp.

[0143] The Discrete Fourier Transform processing unit 903 is used to perform Discrete Fourier Transform processing on all charge and discharge segment data in descending order of the target timestamp to obtain the spectrum diagram corresponding to each charge and discharge segment data.

[0144] The filtering processing unit 904 is used to determine the preset frequency threshold range to which each charge-discharge segment data belongs based on the spectrum diagram, and to use a filter to filter all charge-discharge segment data to obtain multiple target charge-discharge segment data.

[0145] The filtering unit 904 is specifically used for: acquiring multiple preset frequency threshold intervals; acquiring a spectrum of the charge / discharge segment data for each data segment; the spectrum includes multiple frequency components; for each preset frequency threshold interval, calculating the number of frequency components in the spectrum located within the preset frequency threshold interval, and calculating the ratio between the number of frequency components and the total number of frequency components in the spectrum. If the ratio is greater than a preset percentage threshold, it is determined that the charge / discharge segment data belongs to the preset frequency threshold interval, thus obtaining the preset frequency threshold interval to which each charge / discharge segment data belongs.

[0146] The calculation unit 905 is used to mark the target charge-discharge segment data as the first charge-discharge segment data if the preset frequency threshold interval to which any target charge-discharge segment data belongs is a low frequency interval, and to calculate the battery ohmic internal resistance based on the first charge-discharge segment data and the state equation of the battery nth order RC model.

[0147] The calculation unit 905 is specifically used for: obtaining the state equation of the nth-order RC model of the battery; setting the value of n to 1 to simplify the state equation of the nth-order RC model of the battery to obtain the state equation of the pure resistance model; and calculating the ohmic internal resistance of the battery based on the state equation of the pure resistance model and the data of the first charge and discharge segment.

[0148] The estimation unit 906 is used to estimate multiple charging parameters and multiple discharging parameters based on the battery's ohmic internal resistance and all second charge-discharge segment data; the preset frequency threshold range to which the second charge-discharge segment data belongs is a non-low frequency range, and the target timestamp corresponding to the second charge-discharge segment data is greater than the target timestamp corresponding to the first charge-discharge segment data.

[0149] In this embodiment of the invention, the solution obtains battery charging and discharging data within a specific time period from a cloud platform database through a cloud-based BMS, extracts and arranges each charging and discharging segment in chronological order, and adds a timestamp to each data segment. Then, this data is converted from the time domain to the frequency domain to generate a frequency spectrum, and the data is divided according to a set frequency range. Compared with existing cloud-based BMS algorithms, this solution does not rely on large-scale, high-quality training sets and long-term training, and can obtain battery model parameters with high accuracy without neural network algorithms, reducing the demand for cloud computing power and ensuring real-time parameter updates from the cloud-based BMS to the vehicle-side BMS. Simultaneously, frequency analysis can reveal parameter changes in the battery throughout its entire life cycle, meeting the needs of fault diagnosis and safety warning on a long-term scale, and enriching the battery parameter identification methods of cloud-based BMS.

[0150] Combination Figure 9 The device also includes: a first query unit, a second query unit, and a second determination unit, as shown in the diagram.

[0151] The first query unit is used to query whether the first battery data exists in all battery data, and the charging flag in the first battery data is in the on state.

[0152] The second query unit is used to query whether second battery data exists among all battery data if first battery data exists; the timestamp carried by the second battery data is greater than the timestamp carried by the first battery data, and the charging flag in the second battery data is in the off state.

[0153] The second determining unit is used to determine that multiple battery data meet the identification requirements if second battery data exists.

[0154] Combination Figure 9 The first determining unit 902 shown includes: a calculation module, a determining module, and a marking module.

[0155] The calculation module is used to calculate the direction of voltage change gradient based on the voltage and timestamp in each battery data point.

[0156] The determination module is used to determine multiple charge / discharge segment data by combining the current sign, charging flag status, and voltage change gradient direction of each battery data; each charge / discharge segment data includes multiple battery data.

[0157] The tagging module is used to mark the timestamp with the smallest value among the multiple battery data contained in each charge / discharge segment data as the target timestamp corresponding to that charge / discharge segment data.

[0158] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place 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 creative effort.

[0159] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0160] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying battery model parameters, characterized in that, The method includes: Read multiple battery data entries within a preset duration from the cloud platform database; Check if the first battery data exists in all battery data, where the charging flag is on. If the first battery data exists, then check if the second battery data exists among all battery data; the timestamp carried by the second battery data is greater than the timestamp carried by the first battery data, and the charging flag in the second battery data is in the off state; If a second set of battery data exists, then multiple sets of battery data are determined to meet the identification requirements. If multiple battery data sets meet the identification requirements, then multiple charge / discharge segment data sets are determined based on the voltage, current sign, charging flag status, and timestamp in each battery data set; each charge / discharge segment data set corresponds to a target timestamp. According to the target timestamps in descending order, perform discrete Fourier transform on all charge and discharge segment data respectively to obtain the spectrum diagram corresponding to each charge and discharge segment data; Based on the spectrum diagram, the preset frequency threshold range to which each charge / discharge segment data belongs is determined, and all charge / discharge segment data is filtered using a filter to obtain multiple target charge / discharge segment data. If any of the target charge / discharge segment data belongs to a preset frequency threshold range that is a low frequency range, the target charge / discharge segment data is marked as the first charge / discharge segment data, and the battery ohmic internal resistance is calculated based on the first charge / discharge segment data and the state equation of the battery nth order RC model. Based on the battery's ohmic internal resistance and all second charge / discharge segment data, multiple charging parameters and multiple discharging parameters are estimated; the preset frequency threshold range to which the second charge / discharge segment data belongs is a non-low frequency range, and the target timestamp corresponding to the second charge / discharge segment data is greater than the target timestamp corresponding to the first charge / discharge segment data; wherein, the step of estimating multiple charging parameters and multiple discharging parameters based on the battery's ohmic internal resistance and all second charge / discharge segment data includes: treating the battery's ohmic internal resistance as a known quantity, using a parameter identification algorithm, and based on the different responses of each RC network in the n-RC model to each frequency range, combining all second charge / discharge segment data to estimate the RC network parameters in the battery model that have a high response to a certain frequency range, thereby obtaining multiple charging parameters and multiple discharging parameters.

2. The method according to claim 1, characterized in that, The process involves determining multiple charge / discharge segment data based on the voltage, current sign, charging flag status, and timestamp in each piece of battery data, including: Calculate the direction of voltage change gradient based on the voltage and timestamp in each piece of battery data; Multiple charge / discharge segment data are determined by combining the current sign, charging flag status, and voltage change gradient direction of each battery data entry; each charge / discharge segment data includes multiple battery data entries. For each charge / discharge segment data, the timestamp with the smallest value among the multiple battery data contained in the charge / discharge segment data is marked as the target timestamp corresponding to the charge / discharge segment data.

3. The method according to claim 1, characterized in that, The step of determining the preset frequency threshold range to which each charge / discharge segment data belongs based on the spectrum includes: Obtain multiple preset frequency threshold ranges; For each charge / discharge segment data, a spectrum diagram of the charge / discharge segment data is obtained; the spectrum diagram includes multiple frequency components. For each preset frequency threshold interval, the number of frequency components located in the preset frequency threshold interval in the spectrum is calculated, and the ratio between the number of frequency components and the total number of frequency components in the spectrum is calculated. If the ratio is greater than a preset percentage threshold, then the charge / discharge segment data is determined to belong to the preset frequency threshold range, and the preset frequency threshold range to which each charge / discharge segment data belongs is obtained.

4. The method according to claim 1, characterized in that, The step of calculating the battery's ohmic internal resistance based on the first charge / discharge segment data and the battery's nth-order RC model state equation includes: Obtain the state equation of the nth-order RC model of the battery; By setting the value of n to 1, the state equation of the nth-order RC model of the battery is simplified to obtain the state equation of the pure resistance model. Based on the state equation of the pure resistance model and the data from the first charge / discharge segment, the ohmic internal resistance of the battery is calculated.

5. A device for identifying battery model parameters, characterized in that, The device includes: The reading unit is used to read multiple battery data entries within a preset time period from the cloud platform database; The first query unit is used to query whether there is a first battery data in all battery data, where the charging flag in the first battery data is in the on state. The second query unit is used to query whether second battery data exists among all battery data if first battery data exists; the timestamp carried by the second battery data is greater than the timestamp carried by the first battery data, and the charging flag in the second battery data is in the off state; The second determining unit is used to determine that multiple sets of battery data meet the identification requirements if second battery data exists. The first determining unit is configured to, if multiple pieces of battery data meet the identification requirements, determine multiple charge / discharge segment data based on the voltage, current sign, charging flag status, and timestamp in each piece of battery data; each charge / discharge segment data corresponds to a target timestamp. The Discrete Fourier Transform (DFT) processing unit is used to perform DFT processing on all charge-discharge segment data in descending order of the target timestamps to obtain the spectrum diagram corresponding to each charge-discharge segment data. The filtering processing unit is used to determine the preset frequency threshold range to which each charge-discharge segment data belongs based on the spectrum diagram, and to use a filter to filter all charge-discharge segment data to obtain multiple target charge-discharge segment data. The calculation unit is used to mark the target charge-discharge segment data as the first charge-discharge segment data if the preset frequency threshold interval to which any of the target charge-discharge segment data belongs is a low frequency interval, and to calculate the battery ohmic internal resistance based on the first charge-discharge segment data and the state equation of the battery nth order RC model. An estimation unit is used to estimate multiple charging parameters and multiple discharging parameters based on the battery's ohmic internal resistance and all second charge-discharge segment data; the preset frequency threshold interval to which the second charge-discharge segment data belongs is a non-low frequency interval, and the target timestamp corresponding to the second charge-discharge segment data is greater than the target timestamp corresponding to the first charge-discharge segment data; wherein, the estimation unit is specifically used to: treat the battery's ohmic internal resistance as a known quantity, use a parameter identification algorithm, and based on the different responses of each RC network in the n-RC model to each frequency interval, combine all second charge-discharge segment data to estimate the RC network parameters in the battery model that have a high response to a certain frequency interval, thereby obtaining multiple charging parameters and multiple discharging parameters.

6. The apparatus according to claim 5, characterized in that, The first determining unit includes: The calculation module is used to calculate the direction of voltage change gradient based on the voltage and timestamp in each piece of battery data; The determination module is used to determine multiple charge / discharge segment data by combining the current sign, charging flag status, and voltage change gradient direction of each piece of battery data; each piece of charge / discharge segment data includes multiple pieces of battery data; The marking module is used to mark the timestamp with the smallest value among the multiple battery data contained in each charge / discharge segment data as the target timestamp corresponding to that charge / discharge segment data.

7. The apparatus according to claim 5, characterized in that, The filtering unit is specifically used for: Multiple preset frequency threshold intervals are obtained; for each charge / discharge segment data, a spectrum of the charge / discharge segment data is obtained; the spectrum includes multiple frequency components; for each preset frequency threshold interval, the number of frequency components located in the preset frequency threshold interval in the spectrum is calculated, and the ratio between the number of frequency components and the total number of frequency components in the spectrum is calculated. If the ratio is greater than a preset percentage threshold, then the charge / discharge segment data is determined to belong to the preset frequency threshold range, and the preset frequency threshold range to which each charge / discharge segment data belongs is obtained.

8. The apparatus according to claim 5, characterized in that, The computing unit is specifically used for: Obtain the state equation of the nth-order RC model of the battery; set the value of n to 1, simplify the state equation of the nth-order RC model of the battery to obtain the state equation of the pure resistance model; calculate the ohmic internal resistance of the battery based on the state equation of the pure resistance model and the first charge and discharge segment data.

Images