Apparatus and method for battery state-of-charge estimation

Abstract

An apparatus for battery state-of-charge (SOC) estimation, comprising: a pre-processing unit (200) configured to detect an operating parameter of a battery and to filter at least a portion of the parameter to generate a filtered parameter; an SOC estimation unit (300) configured to calculate a nominal SOC based on the operating parameter and the filtered parameter by using a first neural network; a validity estimation unit (400) configured to estimate a validity value indicating a validity of a nominal SOC output by the SOC estimation unit (300) based on the operating parameter and the filtered parameter; and an SOC Kalman filter (500) configured to execute a Kalman filtering algorithm based on the operating parameter, the transmitted nominal SOC, and the validity value, thereby outputting an estimated SOC of the battery.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This disclosure claims priority to an invention patent application entitled “Device and method for battery charge state estimation” and having patent application number US17 / 750,654 filed with the United States Patent and Trademark Office on May 23, 2022, the disclosure of which is incorporated herein by reference in its entirety. Technical Field

[0003] Embodiments of the present disclosure relate generally to the field of batteries, and more particularly, to an apparatus and method for battery state of charge (SOH) estimation. Background Art

[0004] This section introduces aspects that may facilitate a better understanding of the present disclosure. Therefore, the statements in this section should be understood in this context and should not be understood as an admission of what is in the prior art or what is outside the prior art.

[0005] Secondary (rechargeable) batteries are used in energy storage applications in personal portable electronics, electric vehicles, and power systems. Battery SOC, defined as the percentage of remaining capacity relative to the battery's maximum capacity, is an important value that must be calculated by a battery management system (BMS). The BMS uses SOC to provide an indication of when the battery needs to be charged; to extend battery life by preventing operations that result in overcharging or overdischarging; and, in electric vehicle applications, to provide an indication of driving range. Battery SOC cannot be measured directly, so it must be estimated using available measurement data.

[0006] A popular battery SOC estimation method is coulomb counting with open circuit voltage (OCV) and end-of-charge correction. In this method, the BMS calculates the SOC based on the net charge extracted from the battery during operation by integrating the current extracted from the battery (i.e., counting the coulombs extracted). When the battery voltage reaches equilibrium after a sufficiently long rest period, the equilibrium voltage (OCV) can be used to correct the SOC based on the OCV-SOC relationship. Alternatively, the SOC can be corrected by setting this value to 100% at the end of charging, which occurs when the battery voltage reaches the rated value at the end of a well-defined charging sequence. One advantage of this method is its relative simplicity.

[0007] A large amount of research has been devoted to developing battery SOC estimation methods based on Kalman filter algorithms and used in conjunction with the equivalent circuit model (ECM) of the battery. Researchers have reported methods based on the extended Kalman filter (EKF), the unscented Kalman filter (UKF), and the sigma point Kalman filter (SPKF), collectively referred to here as "xKF". In general, the accuracy of these methods depends on the accuracy of the electrochemical model (ECM) and the corresponding xKF covariance parameters.

[0008] The researchers also reported a battery SOC estimation method using the xKF method combined with a neural network SOC model. The neural network SOC model can be developed efficiently and reliably using a commercial neural network optimization package. Summary of the invention

[0009] The inventors of the present disclosure have discovered that existing battery SOC estimation methods all have certain limitations.

[0010] Coulomb counting is known to be inaccurate under certain conditions. If the initial SOC estimate is incorrect, the SOC estimation error will persist until the conditions for the appropriate OCV or end-of-charge SOC correction are met. In addition, if the current is integrated for a long time without SOC correction, it may cause SOC inaccuracies due to the accumulation of current measurement errors.

[0011] For xKF-based approaches used in conjunction with ECM, significant engineering resources may be required to meet accuracy targets. First, the desired ECM prediction accuracy can be obtained by scheduling ECM parameters based on an operating region, which can be defined using sensed physical quantities such as temperature and current and using calculated values ​​such as SOC. Second, the xKF covariance parameters can be scheduled based on the operating region, given that the ECM prediction accuracy varies with the operating region. In many cases, the process of building such a parameter schedule requires significant engineering development resources.

[0012] As for approaches that use the xKF method in conjunction with a neural network SOC model, since model prediction accuracy may vary with operating region (e.g., it may not be accurate in “flat” voltage regions where SOC does not correlate well with available measurement data), the xKF covariance parameters may be scheduled based on the operating region in order to meet accuracy targets. The development of such a parameter schedule may involve significant engineering resources.

[0013] In general, embodiments of the present disclosure provide an apparatus for battery SOC estimation, a method for battery SOC estimation, and a method for calculating parameters for battery SOC estimation. The present disclosure aims to accurately estimate battery SOC using a systematic and efficient development method.

[0014] In a first aspect, there is provided an apparatus for battery state of charge (SOC) estimation, comprising:

[0015] A preprocessing unit (200) is configured to detect operating parameters of the battery and filter at least part of the operating parameters to generate filtered parameters;

[0016] An SOC estimation unit (300) configured to calculate a nominal SOC based on the operating parameters sent from the preprocessing unit and the filtered parameters by using a first neural network;

[0017] A validity estimation unit (400) configured to estimate a validity value based on the operating parameter sent from the preprocessing unit and the filtered parameter, the validity value indicating the validity of the nominal SOC output by the SOC estimation unit;

[0018] The SOC Kalman filter (500) is configured to execute a Kalman filter algorithm based on the operating parameters sent from the preprocessing unit, the nominal SOC sent from the SOC estimation unit, and the effectiveness value sent from the effectiveness estimation unit, thereby outputting an estimated SOC of the battery.

[0019] In a second aspect, there is provided a method for battery state of charge (SOC) estimation, comprising:

[0020] detecting operating parameters of the battery and filtering at least a portion of the operating parameters to generate filtered parameters;

[0021] calculating a nominal SOC based on the operating parameters and the filtered parameters using a first neural network;

[0022] calculating a validity value based on the operating parameter and the filtering parameter, the validity value indicating the validity of the nominal SOC;

[0023] Based on the operating parameters, the nominal SOC and the effectiveness value, a Kalman filter algorithm is executed to output an estimated SOC of the battery.

[0024] In a third aspect, there is provided a method for calculating parameters for battery state of charge (SOC) estimation, comprising:

[0025] Calculate the nominal SOC based on experimental data;

[0026] Calculate the squared nominal SOC error based on experimental data;

[0027] identifying a subset of data from the data set having a squared nominal SOC error less than a first predetermined value; and

[0028] The variance of the difference between the true SOC and the nominal SOC for the identified data subset is calculated, where the result is the parameter R used in the Kalman filter algorithm.

[0029] According to various embodiments of the present disclosure, a first neural network is used to estimate the nominal SOC, a validity value of the nominal SOC is calculated, and a SOC Kalman filter is used to calculate the estimated SOC based on the nominal SOC and the validity value. Therefore, a systematic and efficient development method can be used to accurately estimate the battery SOC. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] By way of example, the above and other aspects, features and benefits of various embodiments of the present disclosure will become more fully apparent from the following detailed description with reference to the accompanying drawings, in which the same reference numerals or letters are used to represent the same or equivalent elements. The accompanying drawings are shown to facilitate a better understanding of the embodiments of the present disclosure and are not necessarily drawn to scale, wherein:

[0031] Figure 1 is a block diagram schematically showing an apparatus for estimating battery SOC according to a preferred embodiment of the present disclosure.

[0032] Figure 2 is a diagram schematically showing a pre-processing unit according to a preferred embodiment of the present disclosure.

[0033] Figure 3 is a diagram schematically illustrating an SOC estimation unit according to a preferred embodiment of the present disclosure.

[0034] Figure 4 is a diagram schematically showing a validity estimation unit according to a preferred embodiment of the present disclosure.

[0035] Figure 5 is a diagram schematically illustrating a SOC Kalman filter according to a preferred embodiment of the present disclosure.

[0036] Figure 6 is a block diagram schematically illustrating an apparatus for estimating the SOC of a battery according to another embodiment of the present disclosure.

[0037] Figure 7 is a flowchart sequentially showing steps for calculating parameters for SOC estimation according to a preferred embodiment of the present disclosure.

[0038] Figure 8 is a flow chart of a method for battery state of charge estimation according to one embodiment.

[0039] Fig. 9 It is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0040] The present disclosure will now be discussed with reference to several example embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and thus implement the present disclosure, rather than implying any limitation on the scope of the present disclosure.

[0041] As used herein, the terms "first" and "second" refer to different elements. Unless the context clearly indicates otherwise, the singular forms "one" and "an" are also intended to include plural forms. The terms "include", "comprise", "have", "have", "include" and / or "include" used herein specify the presence of the features, elements and / or parts, etc., but do not exclude the presence or addition of one or more other features, elements, parts and / or combinations thereof. The term "based on" should be understood as "based at least in part". The terms "one embodiment" and "embodiment" should be understood as "at least one embodiment". The term "another embodiment" should be understood as "at least one other embodiment". Other explicit and implicit definitions may be included below.

[0042] First aspect of the embodiment

[0043] In a first aspect of an embodiment, an apparatus for battery state of charge (SOC) estimation is provided.

[0044] Figure 1 Schematically shows a block diagram of an apparatus 10 for estimating battery SOC according to an embodiment of the present disclosure. Figure 1 As shown in , the apparatus 10 includes a pre-processing unit 200 , an SOC estimation unit 300 , a validity estimation unit 400 , and a SOC Kalman filter 500 .

[0045] In at least one embodiment, the pre-processing unit 200 is configured to detect operating parameters of the battery 100 (such as Figure 1 ), and filtering at least some of the operating parameters to generate filtered parameters. For example, the operating parameters detected by the preprocessing unit 200 include current, voltage, and temperature; the filtered parameters include filtered current and filtered voltage. However, the embodiment is not limited thereto, and the operating parameters and filtered parameters may be of other types.

[0046] The SOC estimation unit 300 is configured to calculate a nominal battery state of charge SOC based on the operating parameters transmitted from the pre-processing unit 200 and the filtered parameters by using a first neural network.

[0047] The effectiveness estimation unit 400 is configured to estimate an effectiveness value based on the operation parameter and the filtered parameter transmitted from the pre-processing unit 200. The effectiveness value indicates the effectiveness of the nominal SOC output by the SOC estimation unit 300.

[0048] The SOC Kalman filter 500 is configured to execute a Kalman filter algorithm based on the operating parameters (eg, current) sent from the preprocessing unit 200 , the nominal SOC sent from the SOC estimation unit 300 , and the effectiveness value sent from the effectiveness estimation unit 400 , thereby outputting an estimated SOC of the battery 100 .

[0049] According to the first aspect of the present disclosure, a first neural network is used to estimate the nominal SOC, a validity value of the nominal SOC is calculated, and a SOC Kalman filter is used to calculate the estimated SOC based on the nominal SOC and the validity value. Therefore, a systematic and efficient development method can be used to accurately estimate the battery SOC.

[0050] Figure 2 Schematically shows a preprocessing unit 200 according to an embodiment of the present disclosure. Figure 2 As shown in , the pre-processing unit 200 includes a current detection unit 210 , a voltage detection unit 220 , a temperature detection unit 230 , a current low-pass filter 240 and a voltage low-pass filter 250 .

[0051] In at least one embodiment, the method for detecting the current, voltage, and temperature of the battery 100 may refer to related art.

[0052] The current low-pass filter 240 may filter the current by using a time constant of T seconds, thereby outputting the filtered current. The voltage low-pass filter 250 may filter the voltage by using a time constant of T seconds, thereby outputting the filtered voltage.

[0053] In a preferred embodiment, the current low-pass filter 240 and the voltage low-pass filter 250 can be implemented according to the following discretized filter calculation:

[0054]

[0055] The calculation cycle time of the current low-pass filter 240 and / or the voltage low-pass filter 250 is T s seconds, the calculation period is k, and the input signal is u k , the filtered signal is x k , the filtered signal in the next calculation cycle is x k+1 For the current low-pass filter 240, the input signal u k is the current, the filtered signal x k is the filtered current. For the voltage low-pass filter 250, the input signal u k is the voltage, the filtered signal x k is the filtered voltage.

[0056] Figure 3 Schematically shows a SOC estimation unit 300 according to an embodiment of the present disclosure. A first neural network (ie, a SOC neural network) is used in the SOC estimation unit 300. Figure 3 As shown in FIG. 310, 310 is the input layer of the SOC neural network. For example, current, voltage, temperature, filtered current and filtered voltage are input to the input layer 310. 320 is a hidden layer for the SOC neural network. 330 is the output layer of the SOC neural network. The working principle of the SOC neural network can refer to the relevant technology.

[0057] Figure 4 Schematically illustrates the effectiveness estimation unit 400 according to one embodiment of the present disclosure. Figure 4 As shown in , the effectiveness estimation unit 400 includes a squared SOC error estimation unit 410 and an effectiveness classification circuit 450 .

[0058] The squared SOC error estimation unit 410 may estimate the squared nominal SOC error by using a second neural network (ie, a squared SOC error neural network) based on the operating parameters and the filtered parameters. The squared nominal SOC error indicates the square of the difference between the true SOC and the nominal SOC.

[0059] like Figure 4 As shown in , 420 is the input layer of the second neural network (i.e., the squared SOC error neural network), for example, current, voltage, temperature, filtered current and filtered voltage are input to the input layer 420. 430 is the hidden layer of the squared SOC error neural network. 440 is the output layer of the squared SOC error neural network, for example, the squared nominal SOC error is output from the output layer 440. The working principle of the squared SOC error neural network can refer to the relevant technology.

[0060] like Figure 4 As shown in , the effectiveness classification circuit 450 can output the effectiveness value based on the magnitude of the squared nominal SOC error being lower than a first predetermined value and the execution time elapsed since the current low-pass filter initialization and / or the voltage low-pass filter initialization being greater than a second predetermined value. For example, the first predetermined value is E 2 In a preferred embodiment, the parameter E is 1%. The second predetermined value may be 3T, where T is a time constant of the current low-pass filter 240 and the voltage low-pass filter 250 .

[0061] For example, if the squared nominal SOC error is less than E 2 If the execution time since the current low-pass filter initialization and / or the voltage low-pass filter initialization is greater than 3T, the nominal SOC validity is set to 1 (valid). Otherwise, the nominal SOC validity is set to 0 (invalid).

[0062] like Figure 4 As shown in , in a preferred embodiment, the validity classification circuit 450 includes a first judgment unit 451, a second judgment unit 452 and a logic unit 453. The first judgment unit 451 can determine whether the squared nominal SOC error is lower than a first predetermined value. For example, if the squared nominal SOC error is lower than the first predetermined value, the first judgment unit 451 can output 1. The second judgment unit 452 can determine whether the execution time since the initialization of the current low-pass filter and / or the voltage low-pass filter is greater than a second predetermined value. For example, if the execution time since the initialization of the current low-pass filter and / or the voltage low-pass filter is greater than the second predetermined value, the second judgment unit 452 can output 1. The logic unit 453 can output the validity value according to the determination result of the first judgment unit 451 and the determination result of the second judgment unit 452. For example, the logic unit 453 is an AND gate. If both the first judgment unit 451 and the second judgment unit 452 output 1, the logic unit 453 can output 1. However, the embodiment is not limited thereto, and the validity classification circuit 450 can have other structures.

[0063] Figure 5 Schematically illustrates a SOC Kalman filter 500 according to an embodiment of the present disclosure. The SOC Kalman filter 500 includes a gain calculator 510 and an estimated SOC calculator 520 .

[0064] The gain calculator 510 may apply the gain selection unit 560 to calculate the gain value based on the effectiveness value, the parameter Q ( 530 ), and the parameter R ( 540 ).

[0065] In a preferred embodiment of the gain selection unit 560, if the validity value indicates that the nominal SOC is valid (eg, the validity value is 1), the gain value is calculated according to the Kalman filter formula:

[0066]

[0067] The gain is L k , the predicted SOC covariance is And parameter R(540) is R.

[0068] In a preferred embodiment of the gain selection unit 560, if the validity value indicates that the nominal SOC is invalid (e.g., the validity value is 0), the gain is set to 0. Mathematically, this is equivalent to setting the covariance associated with the error between the nominal SOC and the true SOC to infinity when the nominal SOC is invalid.

[0069] If the SOC is not otherwise corrected, the parameter Q (530) is the covariance associated with the rate of increase of the SOC covariance. The parameter Q can be estimated based on a combination of factors such as current sensor accuracy, Coulomb efficiency uncertainty, SOC Kalman filter calculation cycle time uncertainty, and battery capacity uncertainty.

[0070] Parameter R (540) is the covariance associated with the error between the nominal SOC and the true SOC when the validity value indicates that the nominal SOC is valid (eg, the validity value is 1).

[0071] The estimated SOC calculator 520 may calculate the estimated SOC based on the nominal SOC, the current, the parameter B ( 550 ), and the gain value.

[0072] Parameter B refers to the predicted change in SOC at one ampere current during the period from the previous calculation cycle of the SOC Kalman filter to the current calculation cycle of the SOC Kalman filter. Parameter B can be calculated based on the SOC Kalman filter calculation cycle time (T s ) and the rated capacity (C) of the battery 100. In a preferred embodiment, the parameter B (550) is calculated according to the following formula:

[0073]

[0074] exist Figure 1 In the apparatus 10, the parameters used can be preset so that the apparatus 10 can directly use the parameters. The parameters may include the weight of the first neural network, the weight of the second neural network, and the parameter R, etc. The parameters may be calculated by a calculator independent of the apparatus 10.

[0075] like Figure 6 As shown in , in at least another embodiment, the device 10a may include a preprocessing unit 200, an SOC estimation unit 300, an effectiveness estimation unit 400, and an SOC Kalman filter 500. The device 10a also includes a parameter calculator 600, which can calculate parameters used in the device 10a. The parameters may include the weight of the first neural network, the weight of the second neural network, and the parameter R, etc. The same parts of the device 10a and the device 10 will not be described, but the description of the parameter calculator 600 will be described below.

[0076] Figure 7 is a flow chart sequentially illustrating blocks for calculating parameters for SOC estimation by the parameter calculator 600 according to one embodiment of the present disclosure.

[0077] In block S010 , current, voltage, and temperature data are extracted from experimental results of battery testing (ie, experimental data) designed to represent operating conditions of a battery application.

[0078] In block S020, the current is filtered using a low pass filter in a manner consistent with the operation of the current low pass filter 240, the voltage is filtered in a manner consistent with the operation of the voltage low pass filter 250, and the true SOC is calculated using coulomb counting. For example, the parameter calculator 600 may control the pre-processing unit (200) to filter the current and voltage, or the parameter calculator 600 may filter the current and voltage by itself.

[0079] In block S030, the first neural network is trained. For example, a training algorithm such as the Levenberg-Marquardt algorithm is used to optimize the SOC neural network (first neural network) weights, with the goal of minimizing the root mean square (RMS) error between the true SOC value calculated in block S020 and the nominal SOC value calculated by the SOC neural network (first neural network) in training when evaluated at the corresponding current, voltage, temperature, filtered current, and filtered voltage values ​​collected in blocks S010 and S020.

[0080] In block S040, the nominal SOC value is calculated by evaluating the trained SOC neural network at the current, voltage, temperature, filtered current, and filtered voltage values ​​collected in blocks S010 and S020. For example, the parameter calculator 600 controls the SOC estimation unit 300 to calculate the nominal SOC by using the trained SOC neural network (first neural network).

[0081] In block S050 , the squared SOC prediction error is calculated as the square of the difference between the true SOC value and the corresponding nominal SOC value.

[0082] In block S060, the second neural network is trained. For example, a training algorithm such as the Levenberg-Marquardt algorithm is used to optimize the squared SOC error neural network (second neural network) weights, with the goal of minimizing the root mean square (RMS) error between the squared SOC prediction error calculated in block S050 and the nominal SOC error value calculated by the squared SOC error neural network (second neural network) in training when evaluated at the corresponding current, voltage, temperature, filtered current, and filtered voltage values ​​collected in blocks S010 and S020.

[0083] In block S070, a square nominal SOC error value is calculated by evaluating the trained square SOC error neural network at the current, voltage, temperature, filtered current, and filtered voltage values ​​collected in blocks S010 and S020. For example, the parameter calculator 600 controls the effectiveness estimation unit 400 to calculate the square nominal SOC error by using the trained square SOC error neural network (second neural network).

[0084] In block S080, the squared nominal SOC error value of the original data set is identified to be less than E 2 In one exemplary embodiment, parameter E is selected to be 1%.

[0085] In block S090 , for the data subset identified in block S080 , a parameter R is calculated as the variance of the difference between the true SOC value and the corresponding nominal SOC value. The parameter R is used in a Kalman filter algorithm.

[0086] exist Figure 7 In the figure, blocks S010, S020 and S030 refer to the process of calculating the weights of the first neural network; blocks S040, S050 and S060 refer to the process of calculating the weights of the second neural network; blocks S070, S080 and S090 refer to the process of calculating the parameter R.

[0087] In at least one embodiment, Figure 7 The method shown in can be used to calculate the parameters, and the parameters are used in the method for estimating the battery SOC performed by the device 10 or 10a.

[0088] In at least another embodiment, Figure 7 The method shown in and the method for estimating the battery SOC performed by the device 10 or 10a can be implemented independently. Figure 7 , and these parameters are used in a method different from the method implemented by the device 10 or 10a for estimating the battery SOC. For another example, the parameters used by the device 10 or 10a can be calculated by Figure 7 The methods shown in the figure are different ways to calculate them.

[0089] As can be seen from the above embodiments, the first neural network is used to estimate the nominal SOC, the validity value of the nominal SOC is calculated, and the SOC Kalman filter is used to calculate the estimated SOC based on the nominal SOC and the validity value. Therefore, a systematic and efficient development method can be used to accurately estimate the battery SOC.

[0090] In the present disclosure, a second neural network is used to estimate the validity of the nominal SOC based on the operating region. In the operating region where the nominal SOC is predicted to be invalid, the KF gain is set to zero, otherwise the KF gain is calculated according to the KF gain formula.

[0091] The present disclosure provides a method for calculating parameters for battery SOC estimation. The method uses a commercially available neural network optimization package that enables the neural network to have a high degree of accuracy while requiring relatively few engineering resources.

[0092] The present disclosure is found to be particularly helpful in providing accurate SOC estimates for applications where the correlation between SOC and available measurements varies greatly from one operating region to another, such as LFP battery chemistries.

[0093] Second aspect of the embodiment

[0094] In a second aspect of the embodiment, a method for estimating a battery state of charge (SOC) is provided. The method corresponds to the device for estimating a battery state of charge (SOC) provided in the first aspect of the embodiment. The same contents as those in the first embodiment are omitted here.

[0095] Figure 8 is a flow chart of a method for battery state of charge estimation according to one embodiment.

[0096] like Figure 8 As shown in , the method includes:

[0097] S801. Detecting operating parameters of the battery, and filtering at least a portion of the operating parameters to generate filtered parameters;

[0098] S802. Calculate the nominal SOC based on the operating parameters and the filtered parameters by using the first neural network;

[0099] S803. Calculating a validity value based on the operating parameters and the filtered parameters, the validity value indicating the validity of the nominal SOC; and

[0100] S804. Based on the operating parameters, the nominal SOC and the effectiveness value, a Kalman filter algorithm is executed to output an estimated SOC of the battery.

[0101] For details of each block, reference may be made to the corresponding description in the first aspect of the embodiment.

[0102] like Figure 8 As shown in , the method may further include:

[0103] S805. Calculate parameters for battery SOC estimation.

[0104] For details of block S805, please refer to the first aspect of the embodiment. Figure 7 Flowchart of the process.

[0105] As can be seen from the above embodiments, the first neural network is used to estimate the nominal SOC, the second neural network is used to estimate the validity value of the nominal SOC, and the SOC Kalman filter calculates the estimated SOC based on the nominal SOC and the validity value. In addition, the parameters are calculated using a well-defined, optimization-based workflow. Therefore, a systematic and efficient development method can be used to accurately estimate the battery SOC.

[0106] The third aspect of the embodiment

[0107] A third aspect of an embodiment of the present disclosure provides an electronic device, which includes the apparatus 10, 10a or the parameter calculator 600 described in the first aspect of the embodiment.

[0108] Fig. 9 Schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Fig. 9 As shown in FIG. 1 , the electronic device 900 may include a processor 910 and a memory 920, and the memory 920 is coupled to the processor 910. The memory 920 may store various data, and may also store a program 930 for data processing, and execute the program 930 under the control of the processor 910.

[0109] In one implementation, the functionality of the apparatus 10, 10a or the parameter calculator 600 may be integrated into the processor 910. The processor 910 may be configured to be able to execute the method in the second aspect of the embodiment.

[0110] In another implementation, the device 10, 10a or parameter calculator 600 and the processor 910 may be configured separately; for example, the device 10, 10a or parameter calculator 600 may be configured as a chip connected to the processor 910 and perform the functions of the device 10 or parameter calculator 600 under the control of the processor 910.

[0111] The embodiments of the present disclosure may also provide a computer-readable program, which, when executed in a device or electronic device, enables the device or electronic device to perform the method as described in the second aspect of the embodiments of the present disclosure.

[0112] An embodiment of the present disclosure may also provide a computer storage medium, including a computer-readable program, which enables a device or an electronic device to execute the method as described in the second aspect of the embodiment of the present disclosure.

[0113] The above-mentioned apparatus and method of the present disclosure can be implemented by hardware, or by a combination of hardware and software. The present disclosure relates to such a computer-readable program, when the program is executed by a logic device, the logic device can execute the above-mentioned apparatus or component, or execute the above-mentioned method or step. The present disclosure also relates to a storage medium for storing the above-mentioned program, such as a hard disk, a floppy disk, a CD, a DVD and a flash memory.

[0114] The method / device described with reference to the embodiments of the present disclosure may be directly embodied as hardware, a software module executed by a processor, or a combination thereof. For example, one or more functional block diagrams and / or one or more combinations of functional block diagrams shown in the accompanying drawings may correspond to software modules of a computer program process, or to hardware modules. Such software modules may correspond to the steps shown in the accompanying drawings, respectively. And the hardware module may be implemented, for example, by using a field programmable gate array (FPGA) to fix the soft module.

[0115] The soft module may be located in RAM, flash memory, ROM, EPROM and EEPROM, registers, hard disk, floppy disk, CD-ROM or any other form of storage medium known in the art. The storage medium may be coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium; or the storage medium may be a component of the processor. The processor and the storage medium may be located in an ASIC. The soft module may be stored in the memory of the mobile terminal or in a memory card of a pluggable mobile terminal. For example, if the device (such as a mobile terminal) adopts a relatively large-capacity MEGA-SIM card or a large-capacity flash memory device, the soft module may be stored in the MEGA-SIM card or the large-capacity flash memory device.

[0116] One or more functional blocks and / or one or more combinations of functional blocks in the accompanying drawings may be implemented as a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, or any suitable combination thereof for performing the functions described in the present application. And one or more functional block diagrams and / or one or more combinations of functional block diagrams in the accompanying drawings may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple processors, one or more microprocessors in communication with a DSP, or any other such configuration.

[0117] The present disclosure has been described above with reference to specific embodiments. However, it should be understood by those skilled in the art that such description is illustrative only and is not intended to limit the scope of protection of the present invention. According to the spirit and principles of the present invention, those skilled in the art may make various modifications and variations, and these modifications and variations fall within the scope of the present invention.

Claims

1. A device for estimating a battery state of charge, characterized in that: include: A preprocessing unit (200) is configured to detect operating parameters of the battery and filter at least part of the operating parameters to generate filtered parameters; an SOC estimation unit (300) configured to calculate a nominal battery state of charge SOC based on the operating parameters sent from the pre-processing unit (200) and the filtered parameters by using a first neural network; a validity estimation unit (400) configured to estimate a validity value based on the operating parameter sent from the preprocessing unit (200) and the filtered parameter, the validity value indicating the validity of the nominal SOC output by the SOC estimation unit (300); The SOC Kalman filter (500) is configured to execute a Kalman filter algorithm based on the operating parameters sent from the preprocessing unit (200), the nominal SOC sent from the SOC estimation unit (300) and the effectiveness value sent from the effectiveness estimation unit (400), thereby outputting an estimated SOC of the battery.

2. The device for battery charge state estimation according to claim 1, characterized in that: The operating parameters detected by the pre-processing unit (200) include current, voltage and temperature, The filtered parameters include filtered current and filtered voltage, The operating parameters transmitted to the SOC Kalman filter (500) include the current, The pre-processing unit (200) comprises: a current low-pass filter (240) configured to filter the current using a time constant of T seconds and output the filtered current; and The voltage low-pass filter (250) is configured to filter the voltage with a time constant of T seconds and output the filtered voltage.

3. The device for battery charge state estimation according to claim 2, characterized in that: The effectiveness estimation unit (400) comprises: a squared SOC error estimation unit (410) configured to estimate a squared nominal SOC error based on the operating parameter and the filtered parameter by using a second neural network, wherein the squared nominal SOC error indicates a square of a difference between a true SOC and the nominal SOC; and The effectiveness classification circuit (450) is configured to output the effectiveness value based on the magnitude of the squared nominal SOC error and the execution time that has elapsed since the current low-pass filter was initialized and / or the voltage low-pass filter was initialized.

4. The device for battery charge state estimation according to claim 3, characterized in that: When the magnitude of the squared nominal SOC error is lower than a first predetermined value (E 2 ) and the execution time elapsed since the current low-pass filter was initialized and / or the voltage low-pass filter was initialized is greater than a second predetermined value (3T), the validity value indicates that the nominal SOC is valid.

5. The device for battery charge state estimation according to claim 4, characterized in that: The validity classification circuit (450) comprises: A first determination unit (451) is configured to determine whether the squared nominal SOC error is lower than the first predetermined value; a second determination unit (452) configured to determine whether the execution time elapsed since the current low-pass filter was initialized and / or the voltage low-pass filter was initialized is greater than the second predetermined value; and The logic unit (453) is configured to output the validity value according to the determination result of the first judgment unit (451) and the determination result of the second judgment unit (452).

6. The device for battery charge state estimation according to claim 2, characterized in that: The SOC Kalman filter (500) comprises: a gain calculator (510) configured to calculate a gain value based on the validity value, a parameter Q and a parameter R, wherein the parameter Q is a covariance associated with a rate of increase of a SOC covariance if no other correction is made to the SOC, and the parameter R is a covariance associated with an error between the nominal SOC and a true SOC when the validity value indicates that the nominal SOC is valid; and An estimated SOC calculator (520) is configured to calculate the estimated SOC based on the nominal SOC, the current, a parameter B and the gain value, wherein the parameter B is calculated according to a SOC Kalman filter cycle time (T s ) and the rated capacity (C) of the battery.

7. The device for battery charge state estimation according to claim 2, characterized in that: The apparatus further comprises a parameter calculator, wherein the parameter calculator is configured to: controlling the SOC estimation unit (300) to calculate the nominal SOC based on experimental data; controlling the effectiveness estimation unit (400) to calculate the squared nominal SOC error based on the experimental data; identifying a subset of data from the data set where the squared nominal SOC error is less than a first predetermined value; as well as The variance of the difference between the actual SOC of the identified data subset and the nominal SOC is calculated, wherein the calculated result is the parameter R used in the Kalman filter algorithm.

8. The device for battery charge state estimation according to claim 7, characterized in that: The parameter calculator is further configured to: Based on the experimental data, the true SOC was calculated using the coulomb counting method; training the first neural network with the goal of minimizing a root mean square error between the true SOC and the nominal SOC calculated by the first neural network being trained; calculating a squared SOC prediction error as the square of the difference between the true SOC and the nominal SOC calculated by the trained first neural network; as well as The second neural network is trained with the goal of minimizing the root mean square error between the squared SOC prediction error and the squared nominal SOC error calculated by the second neural network being trained.

9. A method for battery state of charge (SOC) estimation, characterized in that: include: detecting operating parameters of the battery, and filtering at least some of the operating parameters to generate filtered parameters; calculating a nominal SOC based on the operating parameter and the filtered parameter using a first neural network; calculating a validity value based on the operating parameter and the filtered parameter, the validity value indicating the validity of the nominal SOC; as well as Based on the operating parameters, the nominal SOC and the effectiveness value, a Kalman filter algorithm is executed by a SOC Kalman (500) filter to output an estimated SOC of the battery.

10. The method for battery state of charge estimation according to claim 9, characterized in that: The operating parameters detected include current, voltage and temperature, The filtered parameters include filtered current and filtered voltage, The operating parameters for executing the Kalman filter algorithm include the current, The filtered current is generated by filtering the current with a current low pass filter using a time constant of T seconds; and The filtered voltage is generated by filtering the voltage with a voltage low-pass filter using a time constant of T seconds.

11. The method for battery state of charge estimation according to claim 10, characterized in that: Calculating the effectiveness value includes: estimating a squared nominal SOC error based on the operating parameter and the filtered parameter using a second neural network, the squared nominal SOC error indicating a square of a difference between a true SOC and the nominal SOC; and The validity value is output based on a magnitude of the squared nominal SOC error and an execution time that has elapsed since the current low-pass filter was initialized and / or the voltage low-pass filter was initialized.

12. The method for battery state of charge estimation according to claim 11, characterized in that: When the magnitude of the squared nominal SOC error is lower than a first predetermined value (E 2 ) and the execution time elapsed since the current low-pass filter was initialized and / or the voltage filter was initialized is greater than a second predetermined value (3T), the validity value indicates that the nominal SOC is valid.

13. The method for battery state of charge estimation according to claim 12, characterized in that: The validity value is determined by the following steps: determining whether the squared nominal SOC error is less than the first predetermined value; determining whether the execution time elapsed since the current low-pass filter was initialized and / or the voltage filter was initialized is greater than the second predetermined value; as well as The validity value is outputted based on a determination of the squared nominal SOC error and a determination of the execution time that has elapsed since the current low-pass filter was initialized and / or the voltage filter was initialized.

14. The method for battery state of charge estimation according to claim 10, characterized in that: The estimated SOC is calculated by the following steps: calculating a gain value based on the validity value, a parameter Q, and a parameter R, wherein the parameter Q is a covariance associated with a rate of increase of a SOC covariance if no other correction is made to the SOC, and the parameter R is a covariance associated with an error between the nominal SOC and a true SOC when the validity value indicates that the nominal SOC is valid; as well as The estimated SOC is calculated based on the nominal SOC, the current, a parameter B and the gain value, wherein the parameter B is calculated according to a SOC Kalman filter cycle time (T s ) and the rated capacity (C) of the battery.

15. The method for battery state of charge estimation according to claim 10, characterized in that: The method further comprises: Calculating the nominal SOC based on experimental data; calculating the squared nominal SOC error based on the experimental data; identifying a subset of data from the data set where the squared nominal SOC error is less than a first predetermined value; and The variance of the difference between the actual SOC and the nominal SOC of the identified data subset is calculated, wherein the calculated result is the parameter R used in the Kalman filter algorithm.

16. The method for battery state of charge estimation according to claim 10, characterized in that: The method further comprises: Based on experimental data, the true SOC was calculated using the coulomb counting method; Training a first neural network with the goal of minimizing a root mean square error between the true SOC and a nominal SOC calculated by the first neural network being trained; calculating a squared SOC prediction error as the square of the difference between the true SOC and the nominal SOC calculated by the trained first neural network; and The second neural network is trained with the goal of minimizing the root mean square error between the squared SOC prediction error and the squared nominal SOC error calculated by the second neural network under training.

Images