Apparatus and method for estimating the battery SOC

Abstract

A device for estimating the state of charge (SOC) of a battery, comprising: a preprocessing unit (200), configured to detect the operating parameters of a battery and filter at least a portion of the parameters to generate filtered parameters; an SOC estimation unit (300), configured to calculate the nominal SOC using a first neural network based on the operating parameters and the filtered parameters; a validity estimation unit (400), configured to estimate a validity value based on the operating parameters and the filtered parameters, the validity value indicating the validity of the nominal SOC issued by the SOC estimation unit (300); and an SOC Kalman filter (500), configured to perform a Kalman filtering algorithm based on the operating parameters, the transmitted nominal SOC, and the validity value, thereby issuing an estimated SOC of the battery.

Description

Apparatus and method for estimating the battery SOC Cross-reference to related application The description claims priority to patent application US-17 / 750,654, filed with the United States Patent and Trademark Office on May 23, 2022, and entitled "Apparatus and Method for Batter and SOC Estimation". Technical field The realizations described herein generally relate to the field of batteries, and more particularly, to an apparatus and method for estimating the state of charge (SOC) of a battery. Background This section presents aspects that may facilitate a better understanding of the present description. Consequently, the statements in this section should be read from this perspective and should not be understood as admissions of what is prior art or what is not prior art. Secondary (rechargeable) batteries are used in personal portable electronic devices, electric vehicles, and energy storage applications in power systems. The battery's State of Charge (SOC), defined as the percentage of remaining capacity relative to the battery's maximum capacity, is an important value that the battery management system (BMS) must calculate. The BMS uses the SOC to indicate when a battery needs recharging; to extend battery life by preventing overcharging or over-discharging; and, in electric vehicle applications, to provide an indication of the drive interval. The battery's SOC cannot be measured directly and must therefore be estimated using available measurements. One prevalent method for estimating a battery's State of Charge (SOC) is coulomb counting with open-circuit voltage (OCV) and end-of-charge correction. In this methodology, the Battery Management System (BMS) calculates the SOC based on the net charge drawn from the battery during operation by integrating the current drawn from the battery (i.e., counting the coulombs drawn). 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-to-SOC ratio. Alternatively, the SOC can be corrected by setting the value to 100% at the end of the charge, which occurs when the battery voltage reaches a performance value at the end of a well-defined charging sequence. One benefit of this method is its relative simplicity. A significant body of research has been devoted to developing battery state of charge (SOC) estimation methods based on the Kalman filter algorithm and used in conjunction with an equivalent circuit model (ECM) of the battery. Researchers have described methods based on the extended Kalman filter (EKF), the perfume-free Kalman filter (UKF), and the sigma-point Kalman filter (SPKF), collectively referred to herein as "xKF". As a rule, these methods are accurate to the extent that the ECM and the corresponding xKF covariance parameters are accurate. Researchers have also reported on battery SOC estimation methods that use xKF approaches in conjunction with a neural network SOC model. Neural network SOC models can be developed efficiently and reliably using commercially available neural network optimization packages. Summary The inventor of this description discovered that each of the existing methods for estimating the battery's SOC had some limitation. Regarding the coulomb counting method, it is known to be inaccurate under certain conditions. If the initial estimate of the State of Charge (SOC) is incorrect, the SOC estimation error will persist until the conditions for proper OCV or SOC correction are met at the end of the load. Furthermore, prolonged current integration without SOC correction can lead to SOC inaccuracy due to the accumulation of current measurement errors. Regarding the xKF-based method used in conjunction with an ECM, a substantial amount of engineering resources may be required to meet the accuracy targets. First, the required ECM prediction accuracy can be achieved by programming the ECM parameters based on the operating zone, which can be defined using sensing physical quantities such as temperature and current, and calculated values ​​such as the State of Charge (SOC). Second, in cases where the ECM prediction accuracy is known to vary with the operating zone, the xKF covariance parameters can be programmed according to the operating zone. The process of constructing such parameter programs often requires a substantial amount of engineering development resources. Regarding the method for using xKF approaches in conjunction with a neural network SOC model, since the model's prediction accuracy can vary depending on the operating region (for example, it may be inaccurate in "flat" voltage regions where the SOC is not well correlated with available measurements), the xKF covariance parameters can be programmed according to the operating region to meet accuracy targets. Developing such parameter programs can involve significant engineering resources. In general, the embodiments described herein provide an apparatus for estimating battery SOC, a method for estimating battery SOC, and a method for calculating the parameters used for battery SOC estimation. The aim of this description is to accurately estimate battery SOC using a systematic and efficient development methodology. Firstly, a device is provided for estimating the state of charge (SOC) of the battery, which includes: a preprocessing unit, configured to detect the operating parameters of a battery, and filter at least part of the parameters to generate filtered parameters; a SOC estimation unit, configured to calculate the nominal SOC using a first neural network based on the operating parameters and filtered parameters transmitted to it from the preprocessing unit; a validity estimation unit, configured to estimate a validity value based on the operating parameters and the filtered parameters transmitted to it from the preprocessing unit, the validity value indicating the validity of the nominal SOC issued by the SOC estimation unit (300); A Kalman filter of the SOC, configured to perform a Kalman filtering algorithm based on the operating parameters transmitted to it from the preprocessing unit, the nominal SOC transmitted to it from the SOC estimation unit, and the validity value transmitted to it from the validity estimation unit, thus issuing an estimated battery SOC. Secondly, a method for estimating the state of charge (SOC) of the battery is provided, which includes: detect the operating parameters of a battery, and filter at least part of the parameters to generate the filtered parameters; calculate the nominal SOC using a first neural network based on the operating parameters and the filtered parameters; calculate a validity value based on the operating parameters and the filtered parameters, indicating the validity value as the validity of the nominal SOC; Perform a Kalman filtering algorithm based on the operating parameters, nominal SOC, and validity value, thereby issuing an estimated SOC of the battery. The method may also include calculate a nominal SOC based on experimental data; calculate a squared error of the nominal SOC based on experimental data; identify a subset of data from the dataset for which the root mean square error of the nominal SOC is less than a first predetermined value; and Calculate the variance of the difference between the actual SOC and the nominal SOC for the identified data subset, where the result of the calculation is the parameter R used in the Kalman filtering algorithm. According to various implementations of this description, a neural network is used to estimate the nominal SOC, a validity value for the nominal SOC is calculated, and a Kalman filter of the SOC calculates an estimated SOC based on the nominal SOC and the validity value. Therefore, a systematic and efficient development methodology can be used to accurately estimate the battery's SOC. Brief description of the drawings The aforementioned aspects, characteristics, and benefits, as well as others, of the various embodiments described will become more apparent, by way of example, from the following detailed description with reference to the accompanying drawings, in which similar reference numbers or letters are used to designate similar or equivalent elements. The drawings are illustrated to facilitate a better understanding of the embodiments described and are not necessarily drawn to scale. Figure 1 is a block diagram that schematically illustrates an apparatus for estimating the SOC of a battery according to a preferred embodiment of the present description. Figure 2 is a diagram that schematically illustrates a preprocessing unit according to a preferred embodiment of the present description. Figure 3 is a diagram that schematically illustrates a SOC estimation unit according to a preferred embodiment of the present description. Figure 4 is a diagram that schematically illustrates a validity estimation unit according to a preferred embodiment of the present description. Figure 5 is a diagram that schematically illustrates a Kalman filter of the SOC according to a preferred embodiment of the present description. Figure 6 is a block diagram that schematically illustrates an apparatus for estimating the SOC of a battery according to another embodiment of the present description. Figure 7 is a flowchart that sequentially illustrates the steps for calculating the parameters used for SOC estimation according to a preferred embodiment of the present description. Figure 8 is a flowchart of a method for estimating the state of charge of the battery according to one embodiment. Figure 9 is a schematic diagram of an electronic equipment structure for the implementation of this description. Detailed description The present description will now be analyzed with reference to several illustrative embodiments. It should be understood that these embodiments are analyzed only for the purpose of enabling those skilled in the art to better understand and therefore implement the present description, rather than to suggest any limitation on its scope. The scope of the present invention is defined by the appended claims. As used herein, the terms "first" and "second" refer to different elements. The singular forms "a" and "an" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The terms "comprises," "comprising," "has," "having," "includes," and / or "including," as used herein, specify the presence of the stated and similar features, elements, and / or components, but do not exclude the presence or addition of one or more additional features, elements, components, and / or combinations thereof. The term "based on" should be read as "based at least in part on." The term "an embodiment" should be read as "at least one embodiment." The term "another embodiment" should be read as "at least one other embodiment." Other definitions, both explicit and implicit, may follow. First aspect of the achievements A device for estimating the state of charge (SOC) of the battery is provided in a first aspect of the embodiments. Figure 1 is a block diagram schematically illustrating an apparatus 10 for estimating the battery's SOC according to one embodiment of the present description. As shown in Figure 1, the apparatus 10 includes a preprocessing unit 200, an SOC estimation unit 300, a validity estimation unit 400, and a Kalman filter 500 for the SOC. According to the invention, the preprocessing unit 200 is configured to detect the operating parameters 100 of a battery (as shown in Figure 1) and filter at least some of these 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 to these; the operating parameters and filtered parameters can be of a different type. The SOC estimation unit 300 is configured to calculate the nominal SOC using a first neural network based on the operating parameters and filtered parameters passed to it from the preprocessing unit 200. The validity estimation unit 400 is configured to estimate a validity value based on the operating parameters and filtered parameters transmitted to it from the preprocessing unit 200. The validity value indicates the validity of the nominal SOC issued by the SOC estimation unit 300. The SOC's Kalman filter 500 is configured to perform a Kalman filtering algorithm based on the operating parameters (e.g., current) transmitted to it from the preprocessing unit 200, the nominal SOC transmitted to it from the SOC estimation unit 300, and the validity value transmitted to it from the validity estimation unit 400, thereby outputting an estimated battery SOC 100. According to the invention, the first neural network is used to estimate the nominal SOC, the validity value of the nominal SOC is calculated, and the SOC's Kalman filter calculates the estimated SOC based on the nominal SOC and the validity value. Therefore, a systematic and efficient development methodology can be used to accurately estimate the battery's SOC. Figure 2 is a diagram schematically illustrating the preprocessing unit 200 according to one embodiment of the present description. As shown in Figure 2, the preprocessing unit 200 includes a current sensing unit 210, a voltage sensing unit 220, a temperature sensing unit 230, a current low-pass filter 240, and a voltage low-pass filter 250. In at least one embodiment, the method for detecting the current, voltage, and temperature of battery 100 may refer to the related technique. The 240 current low-pass filter can filter current using a time constant of T seconds, thus outputting the filtered current. The 250 voltage low-pass filter can filter voltage using a time constant of T seconds, thus outputting the filtered voltage. 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: Where the calculation cycle time of the current low-pass filter 240 and / or the voltage low-pass filter 250 is Ts seconds, the calculation cycle is k, the input signal is uk, the filtered signal is xk, and the filtered signal for the next calculation cycle is xk+i. For the current low-pass filter 240, the input signal uk is current, and the filtered signal xk is the filtered current. For the voltage low-pass filter 250, the input signal uk is voltage, and the filtered signal xk is the filtered voltage. Figure 3 is a diagram schematically illustrating the SOC estimation unit 300 according to one implementation of the present description. The first neural network (i.e., the SOC neural network) is used in the SOC estimation unit 300. As shown in Figure 3, 310 is an input layer for 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 an output layer for the SOC neural network. The operating principle of the SOC neural network can be found in the related technique. Figure 4 is a diagram that schematically illustrates a validity estimation unit 400 according to one embodiment of the present description. As shown in Figure 4, the validity estimation unit 400 includes a SOC squared error estimation unit 410 and a validity classification circuit 450. The SOC quadratic error estimation unit 410 can estimate the squared error of the nominal SOC using a second neural network (called the SOC quadratic error neural network) based on the operating parameters and the filtered parameters. The squared error of the nominal SOC indicates the square of the difference between the actual SOC and the nominal SOC. As shown in Figure 4, layer 420 is an input layer for the second neural network (specifically, the SOC squared error neural network). For example, current, voltage, temperature, filtered current, and filtered voltage are input to layer 420. Layer 430 is a hidden layer for the SOC squared error neural network. Layer 440 is an output layer for the SOC squared error neural network. For example, the nominal SOC squared error is output from layer 440. The operating principle of the SOC squared error neural network can be found in the related section. As shown in Figure 4, the 450 validity rating circuit can issue a validity value based on the magnitude of the nominal SOC's root mean square error being below a first default value and the elapsed runtime since the initialization of the current low-pass filter and / or the voltage low-pass filter being greater than a second default value. For example, the first default value is E2. In a preferred embodiment, the parameter E is 1%. The second default value can be 3T, where T is the time constant of the 240 current low-pass filter and the 250 voltage low-pass filter. For example, the nominal validity of the SOC is set to 1 (valid) if the squared error of the nominal SOC is less than E2 and the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter is greater than 3T. Otherwise, the nominal validity of the SOC is set to 0 (invalid). As shown in Figure 4, in a preferred embodiment, the validity classification circuit 450 includes a first evaluation unit 451, a second evaluation unit 452, and a logic unit 453. The first evaluation unit 451 can determine if the squared error of the nominal SOC is below the first default value; for example, the first evaluation unit 451 can output 1 if the squared error of the nominal SOC is below the first default value.The second evaluation unit 452 can determine if the elapsed runtime since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter is greater than the second default value. For example, the second evaluation unit 452 can output 1 if the elapsed runtime since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter is greater than the second default value. Logic unit 453 can output a validity value based on the results of the first evaluation unit 451 and the second evaluation unit 452. For example, logic unit 453 is an AND gate, which can output 1 if both the first evaluation unit 451 and the second evaluation unit 452 output 1.However, the implementation is not limited to these; the 450 validity classification circuit may have a different structure. Figure 5 is a diagram that schematically illustrates a Kalman filter 500 of the SOC according to one embodiment of the present description. The Kalman filter 500 of the SOC includes a gain calculator 510 and an estimated SOC calculator 520. The 510 gain calculator can apply the 560 gain selection unit to calculate the gain value based on the validity value, parameter Q (530), and parameter R (540). In a preferred embodiment of the 560 gain selection unit, if the validity value indicates that the nominal SOC is valid (e.g., the validity value is 1), the gain value is calculated according to the Kalman filter formulation: where the gain is Lk, the predicted SOC covariance is, and the parameter R (540) is R. In a preferred realization of gain, the selection unit 560, if the validity value indicates that the nominal SOC is invalid (e.g., the validity value is 0), then the gain is set to 0. Mathematically, this is equivalent to setting the covariance associated with an error between the nominal SOC and an actual SOC to infinity when the nominal SOC is invalid. The parameter Q (530) is the covariance associated with a gain rate in the covariance of the SOC if the SOC is not corrected in any other way. The parameter Q can be estimated based on a combination of factors such as the accuracy of the current sensor, the uncertainty of the coulombic efficiency, the uncertainty of the calculation cycle time of the SOC's Kalman filter, and the uncertainty of the cell capacity. The parameter R(540) is the covariance associated with an error between the nominal SOC and an actual SOC when the validity value indicates that the nominal SOC is valid (e.g., the validity value is 1). The 520 Estimated SOC Calculator can calculate the estimated SOC based on the nominal SOC value, current, parameter B (550), and gain value. Parameter B is the expected change in the SOC, during the time period from the previous calculation cycle of the SOC Kalman filter to the current calculation cycle of the SOC Kalman filter, for one ampere of current. The parameter B can be determined according to the calculation cycle time of the SOC Kalman filter (Ts) and the performance capacity (C) of the battery 100. In a preferred embodiment, the parameter B (550) is calculated according to the following formula: In Figure 1, the parameters used in device 10 can be predefined so that device 10 can use them directly. These parameters can include the weights of the first neural network, the weights of the second neural network, the parameter R, and so on. The parameters can be calculated using a calculator that is independent of device 10. As shown in Figure 6, an apparatus 10a includes a preprocessing unit 200, a SOC estimation unit 300, a validity estimation unit 400, and a Kalman filter 500 for the SOC. The apparatus 10a further includes a parameter calculator 600, which can calculate the parameters used in the apparatus 10a. The parameters can include the weights of the first neural network, the weights of the second neural network, and the parameter R, etc. The same parts of apparatus 10a and apparatus 10 will not be described here, but the parameter calculator 600 will be described below. Figure 7 is a flowchart that sequentially illustrates blocks for calculating the parameters used for estimating the SOC using the Parameter Calculator 600, according to an implementation of the present description. In block S010, the current, voltage, and temperature data are extracted from experimental results (i.e., experimental data) of battery tests designed to be representative of the operating conditions of the battery application. 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 actual SOC is calculated using the coulomb count. For example, the parameter calculator 600 can control the preprocessing unit 200 to filter the current and voltage, or the parameter calculator 600 can filter the current and voltage itself. In block S030, the first neural network is trained. For example, the weights of the SOC neural network (the first neural network) are optimized using a training algorithm such as the Levenberg-Marquardt algorithm, with the goal of minimizing the RMS error between the actual SOC values ​​calculated in block S020 and the nominal SOC values ​​calculated by the SOC neural network (the first neural network) under training when evaluated against the corresponding values ​​of current, voltage, temperature, filtered current, and filtered voltage collected in blocks S010 and S020. In block S040, the nominal SOC values ​​are calculated by evaluating the trained SOC neural network on the current, voltage, temperature, filtered current, and filtered voltage values ​​collected in blocks S010 and S020. For example, parameter calculator 600 controls SOC estimation unit 300 to calculate the nominal SOC using the trained SOC neural network (first neural network). In block S050, the SOC prediction squared error is calculated as the square of the difference between the actual SOC values ​​and the corresponding nominal SOC values. In block S060, the second neural network is trained. For example, the weights of the SOC squared error neural network (second neural network) are optimized using a training algorithm such as the Levenberg-Marquardt algorithm, with the goal of minimizing the RMS error between the SOC prediction squared error calculated in block S050 and the nominal SOC squared error values ​​calculated by the SOC squared error neural network (second neural network) under training when evaluated to the corresponding values ​​of current, voltage, temperature, filtered current, and filtered voltage collected in blocks S010 and S020. In block S070, the squared values ​​of the nominal SOC are calculated by evaluating the trained SOC squared error neural network on the current, voltage, temperature, filtered current, and filtered voltage values ​​collected in blocks S010 and S020. For example, parameter calculator 600 controls validity estimation unit 400 to calculate the squared error of the nominal SOC using the trained nominal SOC squared error neural network (second neural network). In block S080, a subset of the original dataset is identified for which the nominal SOC root mean square error values ​​are less than E2. In an illustrative embodiment, the parameter E was chosen to be 1%. In block S090, the parameter R is calculated as the variance of the difference between the actual SOC values ​​and the corresponding nominal SOC values, for the subset of data identified in block S080. The parameter R is used in the Kalman filtering algorithm. In Figure 7, blocks S010, S020, and S030 refer to a process for calculating the weights of the first neural network; blocks S040, S050, and S060 refer to a process for calculating the weights of the second neural network; blocks S070, S080, and S090 refer to a process for calculating the parameter R. In at least one embodiment, the method shown in Figure 7 can be implemented to calculate the parameters, and the parameters are used in a method for estimating the battery SOC performed by Apparatus 10 or 10a. In at least one other embodiment, the method shown in Figure 7 and the method for estimating the battery SOC performed by Apparatus 10 or 10a can be implemented independently. For example, the method shown in Figure 7 is provided to calculate the parameters, and these parameters are used in a method different from that implemented by Apparatus 10 or 10a to estimate the battery SOC. In another example, the parameters used by Apparatus 10 or 10a can be calculated using a method different from the one shown in Figure 7. As can be seen in the embodiments mentioned above, the first neural network is used to estimate the nominal SOC, the validity value of the nominal SOC is calculated, and the Kalman filter of the SOC calculates the estimated SOC based on the nominal SOC and the validity value. Therefore, a systematic and efficient development methodology can be used to accurately estimate the battery SOC. In this description, a second neural network is used to estimate the validity of the nominal SOC based on the operating region. The KF gain is set to zero in the operating regions where the nominal SOC is predicted to be invalid, and the KF gain is calculated according to the otherwise KF gain formulation. This description provides a method for calculating the parameters used to estimate the battery's State of Charge (SOC). The method employs commercially available neural network optimization packages, which allow for a high degree of accuracy in the neural networks while requiring relatively few engineering resources. The present description is considered to be especially useful for providing an accurate estimate of the SOC for applications (such as for LFP cell chemicals) where the degree of correlation between the SOC and available measurements varies substantially from one operating area to another. Second aspect of the achievements A method for estimating the state of charge (SOC) of the battery is provided in the second aspect of the embodiments. This method corresponds to the device for estimating the state of charge (SOC) of the battery provided in the first aspect of the embodiments. The same contents as in the first aspect of the embodiments are omitted. Figure 8 is a flowchart of a method for estimating the state of charge of the battery according to one embodiment. As shown in Figure 8, the method includes: S801. The operating parameters of a battery are detected, and at least part of the parameters are used to generate filtered parameters; S802. The nominal SOC is calculated using a first neural network based on the operating parameters and the filtered parameters; S803. A validity value is calculated based on the operating parameters and the filtered parameters, with the validity value indicating the validity of the nominal SOC; and S804. A Kalman filtering algorithm is executed based on the operating parameters, nominal SOC, and validity value, thereby outputting an estimated SOC of the battery. The details of each block can be named in the corresponding description in the first aspect of the realizations. As shown in Figure 8, the method may also include: S805 The parameters used for estimating the battery's SOC are calculated. The details of the S805 block can be referred to in the flowchart of figure 7 in the first aspect of the embodiments. As can be seen in the implementations mentioned above, 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 Kalman filter of the SOC calculates the estimated SOC based on the nominal SOC and the validity value. Furthermore, the parameters are calculated using a well-defined, optimization-based workflow. Therefore, a systematic and efficient development methodology can be used to accurately estimate the battery's SOC. Figure 9 is a schematic diagram of the electronic equipment structure described herein. As shown in Figure 9, an electronic device 900 may include a processor 910 and a memory 920, with the memory 920 coupled to the processor 910. The memory 920 can store various data and, in addition, can store a program 930 for data processing and execute the program 930 under the control of the processor 910. In one implementation, the functions of the apparatus 10, 10a or the parameter calculator 600 can be integrated into the processor 910. The processor 910 can be configured to be able to perform the method in the second aspect of the embodiments. In another implementation, the apparatus 10, 10a or parameter calculator 600 and the processor 910 can be configured separately; for example, the apparatus 10, 10a or parameter calculator 600 can be configured as a chip connected to the processor 910, and the functions of the apparatus 10 or parameter calculator 600 are executed under the control of the processor 910. The embodiments of this description may further provide a computer-readable program, which, when executed on an electronic device or piece of equipment, causes the device or piece of electronic equipment to perform the method as described in the second aspect of the embodiments of this description. The embodiments of this description may further provide a computer storage medium, including a computer-readable program, which causes an electronic apparatus or equipment to perform the method described in the second aspect of the embodiments of this description. The devices and methods described above can be implemented using hardware, or using hardware in combination with software. This description refers to such a computer-readable program that, when executed by a logical device, enables the logical device to fabricate the device or components described above, or to perform the methods or steps described above. This description also refers to a storage medium for storing the above program, such as a hard disk, floppy disk, CD, DVD, flash memory, etc. The methods / apparatus described with reference to the embodiments in this description can be incorporated directly as hardware, software modules 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 drawings may correspond to software modules of computer program procedures or to hardware modules. Such software modules may correspond to the steps shown in the drawings. And the hardware module, for example, may be implemented by reinforcing the software modules using a field-programmable gate array (FPGA). Soft modules can be located in RAM, flash memory, ROM, EPROM, EEPROM, registers, hard drives, floppy disks, CD-ROMs, or any other memory medium known in the art. A memory medium can be coupled to a processor, so that the processor can read information from and write information to the memory medium; or the memory medium can be a component of the processor. The processor and memory medium can be located in an ASIC. Soft modules can be stored in the memory of a mobile terminal and can also be stored on a memory card of a plug-in mobile terminal. For example, if a device (such as a mobile terminal) uses a relatively large-capacity MEGA-SIM card or a high-capacity flash memory device, the soft modules can be stored on the MEGA-SIM card or the high-capacity flash memory device.One or more functional blocks and / or one or more combinations of the functional blocks in the drawings may be realized as a universal processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or any appropriate combination thereof that performs the functions described in this application.And the one or more functional block diagrams and / or one or more combinations of the functional block diagrams of the drawings can also be realized as a combination of computer equipment, such as a combination of a DSP and a microprocessor, multiple processors, one or more microprocessors in combination with communication with a DSP, or any other similar configuration. This description has been previously given with reference to particular embodiments. However, those skilled in the art should understand that such a description is merely illustrative and is not intended to limit the scope of protection of the present invention. Various improvements and modifications to the embodiments described above may be made within the scope of the appended claims.

Claims

1. An apparatus for estimating the state of charge (SOC) of a battery comprising: a preprocessing unit (200) configured to detect the operating parameters of a battery and filter at least some of the operating parameters to generate filtered parameters; an SOC estimating unit (300) configured to calculate a nominal SOC using a first neural network based on the operating parameters and the filtered parameters transmitted thereto from the preprocessing unit (200); characterized in that it further comprises a validity estimating unit (400) configured to estimate a validity value based on the operating parameters and the filtered parameters transmitted thereto from the preprocessing unit (200), the validity value indicating the validity of the nominal SOC emitted by the SOC estimating unit (300); and a Kalman filter (500) of the SOC.configured to perform a Kalman filtering algorithm based on the operating parameters transmitted to it from the preprocessing unit (200), the nominal SOC transmitted to it from the SOC estimation unit (300), and the validity value transmitted to it from the validity estimation unit (400), thereby outputting an estimated SOC of the battery.

2. The apparatus for estimating the state of charge of the battery according to claim 1, wherein the operating parameters detected by the preprocessing unit (200) comprise a current, a voltage, and a temperature; the filtered parameters comprise a filtered current and a filtered voltage; the operating parameters transmitted to the Kalman filter (500) of the SOC comprise the current; the preprocessing 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 a low-pass voltage filter (250), configured to filter the voltage using a time constant of T seconds and output the filtered voltage.

3. The apparatus for estimating the state of charge of the battery according to claim 2, wherein the validity estimation unit (400) comprises: a squared error of SOC estimation unit (410), configured to estimate a squared error of the nominal SOC using a second neural network based on the operating parameters and the filtered parameters, wherein the squared error of the nominal SOC indicates the square of the difference between an actual SOC and the nominal SOC; and a validity classification circuit (450),configured to output the validity value based on the magnitude of the squared error of the nominal SOC and the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter.

4. The battery state-of-charge estimating apparatus according to claim 3, wherein, when the magnitude of the squared error of the nominal SOC is below a first predetermined value (E2) and the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter is greater than a second predetermined value (3T), the validity value indicates that the nominal SOC is valid.

5. The battery state-of-charge estimating apparatus according to claim 4, wherein the validity classification circuit (450) comprises: a first evaluation unit (451),configured to determine whether the root mean square error of the nominal SOC is below the first preset value; a second evaluation unit (452), configured to determine whether the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter is greater than the second preset value; and a logic unit (453), configured to output the validity value based on a determination result from the first evaluation unit (451) and a determination result from the second evaluation unit (452).

6. The apparatus for estimating the battery state of charge according to claim 2, wherein the Kalman filter (500) of the SOC 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 the covariance associated with a gain rate in the covariance of the SOC if the SOC is not otherwise corrected, the parameter R is the covariance associated with an error between the nominal SOC and an actual SOC when the validity value indicates that the nominal SOC is valid; and an estimated SOC calculator (520), 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 determined according to a calculation cycle time (Ts) of the SOC Kalman filter and a performance capacity (C) of the battery.

7. The apparatus for estimating the state of charge of the battery according to claim 2, wherein the apparatus further comprises a parameter calculator,and the parameter calculator is configured to: control the SOC estimation unit (300) to calculate a nominal SOC based on experimental data; control the validity estimation unit (400) to calculate a root mean square error of the nominal SOC based on the experimental data; identify a subset of data from the dataset for which the root mean square error of the nominal SOC is less than a predetermined first value; and calculate the variance of the difference between the actual SOC and the nominal SOC for an identified subset of data, wherein the result of the calculation is a parameter R used in the Kalman filtering algorithm.

8. The apparatus for estimating the battery state of charge according to claim 7, wherein the parameter calculator is further configured to: calculate the actual SOC using the coulomb counting method, based on the experimental data; train a first neural network,with the objective of minimizing the mean squared error between the actual SOC and the nominal SOC calculated by the first neural network in the training phase; calculating the mean squared error of the SOC prediction as the square of the difference between the actual SOC and the nominal SOC calculated by the first trained neural network; and training a second neural network, with the objective of minimizing the mean squared error between the mean squared error of the SOC prediction and the mean squared error of the nominal SOC calculated by the second neural network in training.

9. A method for estimating the state of charge, SOC, of ​​a battery comprising: detecting (S801) the operating parameters of a battery,and filtering at least some of the parameters to generate the filtered parameters; calculating (S802) the nominal SOC using a first neural network based on the operating parameters and the filtered parameters; characterized in that the method further comprises: calculating (S803) a validity value based on the operating parameters and the filtered parameters, the validity value indicating the validity of the nominal SOC; and performing (S804) a Kalman filtering algorithm using a Kalman filter (500) of the SOC based on the operating parameters, the nominal SOC, and the validity value, thereby outputting an estimated SOC of the battery.

10. The method for estimating the state of charge of the battery according to claim 9, wherein the detected operating parameters comprise a current, a voltage, and a temperature, and the filtered parameters comprise a filtered current and a filtered voltage,The operating parameters used to perform the Kalman filtering algorithm comprise the current; the filtered current is generated by filtering the current using a current low-pass filter with a time constant of T seconds; and the filtered voltage is generated by filtering the voltage using a voltage low-pass filter with a time constant of T seconds.

11. The method for estimating the battery state of charge according to claim 10, wherein calculating the validity value comprises: estimating the squared error of the nominal SOC using a second neural network based on the operating parameters and the filtered parameters,The squared error of the nominal SOC indicates the square of the difference between an actual SOC and the nominal SOC; and issue the validity value based on the magnitude of the squared error of the nominal SOC and the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage low-pass filter.

12. The method for estimating the battery state of charge according to claim 11, wherein, when the magnitude of the squared error of the nominal SOC is below a first predetermined value (E2) and the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage filter is greater than a second predetermined value (3T), the validity value indicates that the nominal SOC is valid.

13. The method for estimating the battery state of charge according to claim 12, wherein,The validity value is determined by the following steps: determining if the squared error of the nominal SOC is below the first default value; determining if the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage filter is greater than the second default value; and issuing the validity value based on the result of determining the squared error of the nominal SOC and the result of determining the runtime elapsed since the initialization of the current low-pass filter and / or the initialization of the voltage filter.

14. The method for estimating the battery state of charge according to claim 10, wherein 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 the covariance associated with a gain rate in the SOC covariance if the SOC is not otherwise corrected, the parameter R is the covariance associated with an error between the nominal SOC and an actual SOC when the validity value indicates that the nominal SOC is valid; and calculating the estimated SOC based on the nominal SOC, the current, the parameter B, and the gain value, wherein the parameter B is determined according to a calculation cycle time (Ts) of the SOC Kalman filter and a performance capacity (C) of the battery.

15. The method for estimating the state of charge of the battery according to claim 10, wherein,The method further comprises: calculating a nominal SOC based on experimental data; calculating a root mean square error of the nominal SOC based on the experimental data; identifying a subset of data from the dataset for which the root mean square error of the nominal SOC is less than a predetermined first value; and calculating the variance of the difference between the actual SOC and the nominal SOC for the identified subset of data, wherein the result of the calculation is the parameter R used in the Kalman filtering algorithm; or, wherein the method further comprises: calculating the actual SOC using the coulomb counting method, based on experimental data; training a first neural network,with the objective of minimizing the root mean square error between the actual SOC and the nominal SOC calculated by the first neural network in training; calculating the SOC prediction error as the square of the difference between the actual SOC and the nominal SOC calculated by the first trained neural network; and training a second neural network, with the objective of minimizing the root mean square error between the SOC prediction error and the nominal SOC error calculated by the second neural network in training.

Images