Method and system for calibrating state of charge of battery and battery module
By building neural network models and using real-time working data for calibration, the problem of high complexity of SOC calibration in the prior art is solved, efficient, economical and reliable SOC calibration is achieved, and the safety and user experience of the vehicle are improved.
Patent Information
- Application Number
- CN202311593788.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-24
- Publication Date
- 2025-05-27
AI Technical Summary
The calculation complexity of the calibration method for battery state of charge (SOC) in the prior art is difficult to achieve efficient, economical and reliable calibration.
By obtaining the historical working data of the battery, building a neural network model and inputting the real-time working data into the neural network model for calibration, simplifying the calibration process of SOC.
Improves calibration accuracy of SOC, reduces calibration time, improves user experience, and improves vehicle safety.
Smart Images
Figure CN120039159A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of vehicles, and more particularly to the field of battery systems for new energy vehicles. More specifically, it relates to a method and system for calibrating the state of charge of a battery, a battery module applying the system, a computer device implementing the method, and a computer-readable storage medium executing the method. Background Art
[0002] In recent years, new energy vehicles have become increasingly common in life. In new energy vehicles, lithium batteries are usually used as the power source. Due to the particularity of lithium batteries, it is necessary to avoid overcharging, over-discharging, and overheating of the battery during use. For example, when the battery is over-discharged, it will cause serious damage to the battery, affect the performance of the entire battery module, and even cause an explosion. Usually, a BMS (Battery Management System) is used to control the state of the battery. Among them, the state of the battery includes SOC (State of Charge), SOH (State-of-Health), SOP (State-of-Power), etc. Since the implementation of functions such as SOH, SOP calculation, and charge equalization of the BMS is based on the SOC value, the calibration accuracy of SOC is particularly important.
[0003] SOC usually reflects the ratio of the remaining power of the current battery to the total available power under a certain discharge current, usually expressed as a percentage. SOC can directly affect the voltage and current of the battery. The value range of SOC is 0-1. Usually, when SOC = 0, it means the battery is fully discharged, and when SOC = 1, it means the battery is fully charged.
[0004] Currently, the main methods for calibrating SOC are the ampere-hour integration method, the Kalman filter method, etc. These methods require a high estimation ability of the model and a high measurement ability of the sensor during use, and the calculation complexity is relatively high. Therefore, it is necessary to improve the calibration method of SOC. Summary of the Invention
[0005] The object of the present invention is to solve the problems existing in the above-mentioned prior art, and provide a method for calibrating the state of charge of a battery, so as to improve the calibration method of SOC in a cost-effective and reliable manner.
[0006] To this end, according to one aspect of the present invention, there is provided a method for calibrating the state of charge of a battery, the method comprising: obtaining an original database, the original database including historical working data of the battery; configuring a neural network model based on the original database; obtaining online data, the online data including real-time working data of the battery; and inputting the online data into the neural network model to calibrate the state of charge of the battery.
[0007] According to the above technical concept, the present invention may further include any one or more of the following optional forms.
[0008] In some optional forms, obtaining the original database includes: obtaining data related to the historical behavior of the battery through a second-order resistor-capacitor battery model.
[0009] In some optional forms, the data related to the historical behavior of the battery includes at least one of the ambient temperature, surface temperature, current, initial voltage, and initial battery capacity of the battery.
[0010] In some optional forms, obtaining the original database further includes: obtaining historical state of charge data of the battery through ampere-hour integration.
[0011] In some optional forms, obtaining the original database further includes: obtaining historical initial state of charge data of the battery after a preset rest time through the open-circuit voltage method.
[0012] In some optional forms, the historical working data of the battery and the real-time working data have the same variables.
[0013] In some optional forms, configuring the neural network model based on the original database includes: constructing and training a neural network model with the feature vector sequence of the historical working data in the original database as the input of the neural network model and the calibrated state of charge data as the output of the neural network model.
[0014] In some optional forms, the neural network model is a long short-term memory network model.
[0015] According to another aspect of the present invention, there is provided a system for calibrating the state of charge of a battery, the system comprising: a first acquisition module configured to acquire an original database, the original database including historical working data of the battery; a processing module configured to configure a neural network model based on the original database; a second acquisition module configured to acquire online data, the online data including real-time working data of the battery; and a calibration module configured to input the online data into the neural network model to calibrate the state of charge of the battery.
[0016] In some alternative forms, the first acquisition module is further configured to: acquire data related to the historical behavior of the battery through a second-order resistor-capacitor battery model.
[0017] In some alternative forms, the first acquisition module is further configured to: acquire the historical state of charge data of the battery through ampere-hour integration method.
[0018] In some alternative forms, the first acquisition module is further configured to: acquire the historical initial state of charge data of the battery after standing for a preset time through open-circuit voltage method.
[0019] In some alternative forms, the processing module is further configured to: use the feature vector sequence of the historical working data in the original database as the input of the neural network model and use the calibrated state of charge data as the output of the neural network model to construct and train the neural network model.
[0020] In some alternative forms, the neural network model is a long short-term memory network model.
[0021] According to another aspect of the present invention, there is provided a battery module, including one or more batteries and the system for calibrating the state of charge of the above-mentioned battery.
[0022] According to another aspect of the present invention, there is provided a computer device, which includes a memory, a processor, and instructions stored on the memory and executable by the processor, wherein when the processor executes the instructions, the method for calibrating the state of charge of the above-mentioned battery is implemented.
[0023] According to another aspect of the present invention, there is provided a computer-readable storage medium, which has computer-executable instructions stored thereon, and the computer-executable instructions are used to execute the method for calibrating the state of charge of the above-mentioned battery.
[0024] The present invention calibrates the state of charge of the battery through a neural network model, simplifies the calibration process of the state of charge, improves the calibration accuracy of the state of charge, enables more accurate estimation of the vehicle mileage, further improves the safety of the vehicle, and improves the user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Other features and advantages of the present invention will be better understood through the following alternative embodiments described in detail in conjunction with the drawings, wherein:
[0026] Figure 1 is a flowchart of a method for calibrating the state of charge of a battery according to an embodiment of the present invention;
[0027] Figure 2 is a flowchart of a method for calibrating the state of charge of a battery according to another embodiment of the present invention;
[0028] Figure 3 is a circuit diagram of a second-order RC battery model according to an embodiment of the present invention;
[0029] Figure 4 is a schematic diagram of a system for calibrating the state of charge of a battery according to an embodiment of the present invention; and
[0030] Figure 5 is a schematic diagram of a computer device according to an embodiment of the present invention. Detailed Embodiments
[0031] The implementation and use of the embodiments are discussed in detail below. However, it should be understood that the specific embodiments discussed are merely illustrative of specific ways of implementing and using the present invention, and do not limit the scope of the present invention.
[0032] In addition, the flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of the methods and systems according to various embodiments of the present invention. It should be noted that the functions marked in the blocks may also occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved.
[0033] In new energy vehicles, accurate SOC estimation affects the driving experience of the driver. If the SOC estimation is correct, the driver will more accurately know the remaining vehicle mileage, and the SOC is also a basic input for BMS management. The inventor found that currently, during the calibration of SOC using methods such as Kalman filtering, the computational complexity is relatively high, resulting in a relatively long calibration time for SOC.
[0034] Referring to Figure 1 , Figure 1 shows a flowchart of a method for calibrating the state of charge of a battery according to an embodiment of the present invention.
[0035] A method for calibrating the state of charge of a battery according to an embodiment of the present invention may include the following steps.
[0036] Step S110: Obtain an original database for constructing a neural network model, where the original database may include historical working data of the battery.
[0037] Step S120: Configure a neural network model based on the obtained original database.
[0038] Step S130: Obtain online data related to the real-time working data of the battery through the BMS.
[0039] Step S140: Input the acquired online data into the configured neural network model to calibrate the state of charge of the battery.
[0040] In this way, according to the SOC calibration method of the present invention, the calibration process of SOC can be simplified and the calibration accuracy of SOC can be improved by using a neural network model. The configured neural network model can achieve a simple estimation of SOC, reduce the calibration time of SOC, and improve the user experience.
[0041] In some embodiments, the historical working data and the real-time working data of the battery may have the same variables, that is to say, the historical working data and the real-time working data of the battery contain the same measured variables. In this way, consistent information can be provided to the neural network model to ensure that the data used in the calibration process is comparable, making it easy for the neural network model to be trained and tested between the historical working data and the real-time working data, thereby improving the reliability and accuracy of SOC calibration.
[0042] In some embodiments, the neural network model can be a Long Short-Term Memory (LSTM) network model. The LSTM model can solve the problem of gradient disappearance in traditional Recurrent Neural Networks (RNNs), so it can better process long sequences, which is beneficial for the input data with variable lengths that may occur in SOC calibration. The LSTM model can flexibly adjust its internal state to adapt to different input sequence lengths, thereby improving the calibration accuracy of SOC. It can be understood that the neural network model is not limited to this, and other neural network models can be selected depending on different needs.
[0043] Refer to Figure 2 , Figure 2 is a flowchart of a method for calibrating the state of charge of a battery according to another embodiment of the present invention.
[0044] The method for calibrating the state of charge of a battery according to another embodiment of the present invention may include the following steps.
[0045] Step S211: Obtain data related to the historical behavior of the battery through a second-order resistor-capacitor battery model (referred to as a second-order RC battery model).
[0046] The second-order RC battery model used in this embodiment can be a second-order RC equivalent circuit model as shown in Figure 3 , which is composed of two RC network structures and a resistor in series. In Figure 3 , U OCrepresents the open-circuit voltage of the battery, R represents the internal resistance of the battery, R 1 represents the electrochemical polarization resistance, R 2 represents the concentration polarization resistance, C 1 represents the electrochemical polarization capacitance, C 2 represents the concentration polarization capacitance, i represents the load current, U L represents the port voltage of the battery that can be directly measured.
[0047] Through this second-order RC battery model, data related to the historical behavior of the battery for configuring the neural network model as listed in Table 1 can be obtained.
[0048] Table 1 Data related to the historical behavior of the battery obtained through the second-order RC battery model
[0049]
[0050]
[0051] Step S212: Obtain the historical state of charge data of the battery through the ampere-hour integration method.
[0052] In step S212, the historical state of charge data of the battery can be obtained through the ampere-hour integration method (also known as the Ah integration method). If the SOC at the start state of charge and discharge is denoted as SOC ini , then the SOC at the current state is:
[0053]
[0054] where C is the rated capacity of the battery; I is the battery current.
[0055] Step S213: Obtain the historical initial state of charge data of the battery after standing for a preset time through the open-circuit voltage method.
[0056] In step S213, the initial state of charge data of the battery after standing for a preset time can be obtained through the open-circuit voltage method (OCV-SOC curve), that is, the initial state of charge data of the battery can be obtained through the open-circuit voltage method after the vehicle has stopped running for a preset time. In some embodiments, the preset time can be a time greater than 1 hour.
[0057] Generally, the ampere-hour integration method is used to track the change of the battery SOC. In the case where the initial SOC is unknown, the ampere-hour integration method may not be able to provide an accurate initial SOC. Therefore, according to the embodiments of the present invention, the open-circuit voltage method is used to obtain the initial SOC of the battery. Since the open-circuit voltage changes significantly at the initial and end stages of charge and discharge of the battery, the open-circuit voltage method can obtain a relatively accurate initial SOC.
[0058] Step S220: Construct and train a neural network model with the feature vector sequence of historical working data in the original database as the input of the neural network model and the calibrated state of charge data as the output of the neural network model.
[0059] After obtaining the above historical working data, the construction and training of the neural network model can be started. In some embodiments, the feature vector sequence of historical working data in the original database can be used as the input of the neural network model. This feature vector sequence reflects the state of the battery system at different time points. The neural network model can learn the relationships between different time points through the patterns and trends in the feature vector sequence, so as to more accurately capture the state evolution of the battery system. In addition, by training the neural network model, the relationship between historical working data and state of charge data can be understood, which can be used for the optimization of the battery system, etc.
[0060] Step S230: Obtain online data related to the real-time working data of the battery through the BMS.
[0061] Step S240: Input the online data into the configured neural network model to calibrate the state of charge of the battery.
[0062] The configured neural network model can be used in a real-time decision support system. By inputting the current real-time working data, the model can generate state of charge data corresponding to the real-time working data. Based on the generated state of charge data, the state of charge data can be updated in the BMS to complete the calibration process of the state of charge data. And between the calibration processes of the state of charge data in the next cycle, this calibration result can be used as the calculation benchmark for the state of charge data.
[0063] In this way, the accuracy of the state of charge calibration of the battery can be further improved, enabling a more accurate estimation of the vehicle mileage, further enhancing the safety of the vehicle, and improving the user experience.
[0064] Refer to Figure 4 , Figure 4 which shows a schematic diagram of a system for calibrating the state of charge of a battery according to an embodiment of the present invention.
[0065] The system for calibrating the state of charge of a battery according to this embodiment includes a first acquisition module 110, a processing module 120, a second acquisition module 130, and a calibration module 140. Among them, the first acquisition module 110 is used to acquire an original database including the historical working data of the battery. The processing module 120 configures a neural network model based on this original database. The second acquisition module 130 is used to acquire online data including the real-time working data of the battery. The calibration module 140 is used to input this online data into the neural network model to calibrate the state of charge of the battery.
[0066] In this way, the system for calibrating the state of charge of the battery according to this embodiment calibrates the state of charge of the battery through a neural network model, improving the calibration accuracy and simplifying the calibration process.
[0067] In some embodiments, the first acquisition module 110 can acquire data related to the historical behavior of the battery through a second-order RC battery model. In addition, the first acquisition module 110 can also acquire the historical state of charge data of the battery through the ampere-hour integration method and acquire the historical initial state of charge data of the battery after standing for a preset time through the open-circuit voltage method. In this way, more complete and accurate historical working data can be obtained, improving the calibration accuracy of the state of charge.
[0068] In some embodiments, the processing module 120 can construct and train a neural network model with the feature vector sequence of the historical working data in the original database as the input of the neural network model and the calibrated state of charge data as the output of the neural network model, so that the configured neural network model can more accurately understand the relationship between the historical working data and the state of charge data, improving the calibration accuracy.
[0069] Refer to Figure 5 , Figure 5 FIG. is a schematic diagram of a computer device according to an embodiment of the present invention.
[0070] The present invention also provides a computer device 200. As Figure 5 shown, the computer device 200 may include a memory 210 and a processor 220. Instructions 211 may be stored in the memory 210, and the instructions 211 may be executed by the processor 220. Among them, when the processor 220 executes the instructions 211, it implements the method for calibrating the state of charge of the battery according to the above embodiment.
[0071] The computer device 200 of this embodiment may be a device such as a notebook, a desktop computer, and a cloud server. It can be understood that the components included in the computer device 200 are not limited to the memory 210 and the processor 220 and may vary depending on different needs. Exemplarily, the computer device 200 may further include multiple components connected to its input / output interface (Figure 5 not shown in the figure), including but not limited to: an input unit such as a keyboard, a mouse, etc.; an output unit such as a display, a speaker, etc.; a storage unit such as a semiconductor storage device, a magnetic surface storage device, an optical storage device, etc.; and a communication unit such as a network card, a wireless communication transceiver, etc.
[0072] In some embodiments, the memory 210 may include, for example, a Random Access Memory (RAM) or a Read-Only Memory (ROM). The memory 210 may be used to store instructions, programs, codes, and other programs and data required by the computer device 200, but is not limited thereto.
[0073] In addition, the processor 220 may be a Central Processing Unit (CPU), or may be other general-purpose processors such as a Digital Signal Processing (DSP), a Field-Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), etc.
[0074] In an exemplary embodiment of the present invention, there is also provided a computer-readable storage medium having computer-executable instructions stored thereon, and the computer-executable instructions are used to execute the method for calibrating the state of charge of the battery according to the above embodiments.
[0075] Optionally, the computer-readable storage medium according to the present embodiment may be a ROM, a RAM, a semiconductor storage device, a magnetic surface storage device, an optical storage device, etc.
[0076] The method for calibrating the state of charge of the battery of the present invention calibrates the SOC of the battery through a neural network model, simplifies the calibration process of the SOC, improves the calibration accuracy of the SOC, improves the safety performance of the battery, is cost-effective, and has high calibration efficiency. In addition, the calibration method of the present invention is not limited to being used in new energy vehicles, and can also be applied to other types of vehicles such as hybrid vehicles according to needs to calibrate the SOC of the vehicle battery and improve the safety performance of the vehicle. Furthermore, the calibration method of the present invention can be used for various types of lithium batteries, such as lithium iron phosphate batteries, ternary lithium batteries, etc.
[0077] It should be understood that the embodiments shown in the figures only show alternative configuration modes of the system for calibrating the state of charge of the battery according to the present invention. However, they are only illustrative and not restrictive. Other configuration modes can also be adopted without departing from the spirit and scope of the present invention.
[0078] The technical content and technical features of the present invention have been disclosed above. However, it can be understood that under the creative concept of the present invention, those skilled in the art can make various changes and improvements to the above-disclosed concept, but they all fall within the protection scope of the present invention. The description of the above embodiments is illustrative rather than restrictive, and the protection scope of the present invention is determined by the claims.
Claims
1. A method for calibrating the state of charge of a battery, characterized in that, the method includes: Obtain an original database (S110), where the original database includes historical working data of the battery; Configure a neural network model based on the original database (S120); Obtain online data (S130; S230), where the online data includes real-time working data of the battery; Input the online data into the neural network model to calibrate the state of charge of the battery (S140; S240).
2. The method for calibrating the state of charge of a battery according to claim 1, characterized in that, Obtaining the original database includes: obtaining data related to the historical behavior of the battery through a second-order resistor-capacitor battery model (S211).
3. The method for calibrating the state of charge of a battery according to claim 2, characterized in that, The data related to the historical behavior of the battery includes at least one of the ambient temperature, surface temperature, current, initial voltage, and initial battery capacity of the battery.
4. The method for calibrating the state of charge of a battery according to claim 2, characterized in that, Obtaining the original database further includes: obtaining historical state of charge data of the battery through ampere-hour integration method (S212).
5. The method for calibrating the state of charge of a battery according to claim 4, characterized in that, Obtaining the original database further includes: obtaining historical initial state of charge data of the battery after a preset time of standing through the open-circuit voltage method (S213).
6. The method for calibrating the state of charge of a battery according to claim 1, characterized in that, The historical working data of the battery and the real-time working data have the same variables.
7. The method for calibrating the state of charge of a battery according to claim 1, characterized in that, Configuring a neural network model based on the original database includes: constructing and training a neural network model with the feature vector sequence of the historical working data in the original database as the input of the neural network model and the calibrated state of charge data as the output of the neural network model (S220).
8. The method for calibrating the state of charge of a battery according to any one of claims 1 to 7, characterized in that, The neural network model is a long short-term memory network model.
9. A system for calibrating the state of charge of a battery, characterized in that, the system includes: A first acquisition module (110), the first acquisition module (110) is configured to acquire an original database, where the original database includes historical working data of the battery; A processing module (120), the processing module (120) is configured to configure a neural network model based on the original database; A second acquisition module (130), the second acquisition module (130) is configured to acquire online data, where the online data includes real-time working data of the battery; A calibration module (140), the calibration module (140) is configured to input the online data into the neural network model to calibrate the state of charge of the battery.
10. The system for calibrating the state of charge of a battery according to claim 9, Characterized in that, The first acquisition module (110) is further configured to: acquire data related to the historical behavior of the battery through a second-order resistor-capacitor battery model.
11. The system for calibrating the state of charge of the battery according to claim 10, Characterized in that, The first acquisition module (110) is further configured to: acquire the historical state of charge data of the battery through the ampere-hour integration method.
12. The system for calibrating the state of charge of the battery according to claim 11, Characterized in that, The first acquisition module (110) is further configured to: acquire the historical initial state of charge data of the battery after a preset rest time through the open-circuit voltage method.
13. The system for calibrating the state of charge of the battery according to claim 10, Characterized in that, The processing module (120) is further configured to: construct and train a neural network model with the feature vector sequence of the historical working data in the original database as the input of the neural network model and the calibrated state of charge data as the output of the neural network model.
14. The system for calibrating the state of charge of the battery according to any one of claims 9 to 13, Characterized in that, The neural network model is a long short-term memory network model.
15. A battery module, Characterized in that, The battery module includes one or more batteries and the system for calibrating the state of charge of the battery according to any one of claims 9 to 14.
16. A computer device, Characterized in that, The computer device (200) includes a memory (210), a processor (220), and instructions (211) stored on the memory (210) and executable by the processor (220), wherein when the processor (220) executes the instructions (211), the method for calibrating the state of charge of the battery according to any one of claims 1 to 8 is implemented.
17. A computer-readable storage medium, Characterized in that, The computer-readable storage medium has computer-executable instructions stored thereon, and the computer-executable instructions are used to execute the method for calibrating the state of charge of the battery according to any one of claims 1 to 8.