Battery health state estimation method and device, electronic equipment and medium

Abstract

The invention provides a battery health state estimation method and device, electronic equipment and a medium, and the method comprises the steps: obtaining the real-time operation data of a battery, determining the feature data of the battery according to the real-time operation data, and enabling the feature data to represent the aging trend of the battery; inputting the characteristic data into a machine learning model to obtain an initial health state estimation value of the battery, and inputting the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicates a corresponding relationship between the historical feature data of the battery and the health state value of the battery; and inputting the initial health state estimation value and the ohm internal resistance real-time identification value into an adaptive Kalman filter for data fusion to obtain a final health state estimation value of the battery. According to the invention, accurate estimation of the health state of the battery is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of batteries, in particular to a battery state of health estimation method and device, electronic equipment and medium. BACKGROUND

[0002] Battery state of health (SOH) estimation is one of the core functions of a battery management system, and its accuracy directly affects the reliability and safety of device operation. At present, although traditional SOH estimation methods have been widely studied, they all have significant limitations. The method based on voltage change estimates according to the change in voltage characteristics during battery aging. However, the voltage signal is easily disturbed by multiple factors such as temperature and load current, resulting in insufficient estimation accuracy. The internal resistance measurement method estimates using the strong correlation between battery internal resistance and aging degree. Although the mechanism is clear, accurate measurement of internal resistance usually relies on complex equipment, increasing the system cost. SUMMARY

[0003] Therefore, the present application aims to provide a battery state of health estimation method and device, electronic equipment and medium, which overcome at least one of the above-mentioned defects.

[0004] In a first aspect, the present application provides a battery state of health estimation method, which comprises: obtaining real-time operation data of a battery, and determining feature data of the battery according to the real-time operation data, the feature data representing an aging trend of the battery; inputting the feature data into a machine learning model to obtain an initial state of health estimation value of the battery, and inputting the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicating a corresponding relationship between historical feature data of the battery and a state of health value of the battery; inputting the initial state of health estimation value and the ohmic internal resistance real-time identification value into an adaptive Kalman filter for data fusion to obtain a final state of health estimation value of the battery.

[0005] In one possible implementation, the feature data is determined by: obtaining the real-time operation data, the real-time operation data including a voltage sequence, a current sequence and a temperature sequence of the battery in a target charge-discharge cycle; and calculating the feature data according to the voltage sequence, the current sequence and the temperature sequence, the feature data including a constant-current charge voltage platform slope and a discharge curve differential capacity peak value. In one possible implementation, the machine learning model is constructed by: obtaining a training sample set of the battery, the training sample set including historical feature data and measured health state estimates of the battery in multiple charge-discharge cycles; for each training sample in the training sample set, using the historical feature data of the training sample as input and the measured health state estimate of the training sample as output, so as to represent the relationship between the feature data and health state values ​​of the battery using the machine learning model.

[0006] In one possible implementation, the final health state estimate of the battery is obtained by: inputting the initial health state estimate and the real-time ohmic resistance identification value into an adaptive Kalman filter as two observation inputs of the adaptive Kalman filter; dynamically weighing the reliability of the two observation inputs and performing data fusion according to the recursive estimation algorithm of the adaptive Kalman filter to output the final health state estimate.

[0007] In one possible implementation, the adaptive Kalman filter is configured by: defining the battery's health state value as the primary state variable and establishing an observation equation relating the battery's health state value to its ohmic internal resistance; pre-setting a state transition matrix based on a pre-set battery aging empirical model and pre-setting an observation noise covariance matrix based on the measurement noise level, thereby completing the configuration of the adaptive Kalman filter.

[0008] In one possible implementation, the equivalent circuit model is represented by the following formula:

[0009] in, This is the terminal voltage value. Open circuit voltage, For load current, and These are the electrochemical polarization voltage and the concentration polarization voltage, respectively.

[0010] In a second aspect, the present application provides a battery state of health estimation device, the device comprising: an acquisition module configured to acquire real-time operation data of a battery, and determine feature data of the battery according to the real-time operation data, the feature data representing an aging trend of the battery; a prediction module configured to input the feature data into a machine learning model to obtain an initial state of health estimation value of the battery, and input the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicating a corresponding relationship between historical feature data of the battery and a state of health value of the battery; and a correction module configured to input the initial state of health estimation value and the ohmic internal resistance real-time identification value into an adaptive Kalman filter for data fusion to obtain a final state of health estimation value of the battery.

[0011] In a possible implementation, the acquisition module is further configured to: acquire the real-time operation data, the real-time operation data comprising a voltage sequence, a current sequence and a temperature sequence of the battery in a target charge-discharge cycle; and calculate the feature data according to the voltage sequence, the current sequence and the temperature sequence, the feature data comprising a constant-current charge voltage platform slope and a discharge curve differential capacity peak value.

[0012] In a third aspect, the present application further provides an electronic device, comprising: a processor, a memory and a bus, the memory storing machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the machine readable instructions are executed by the processor to perform the steps of the above method.

[0013] In a fourth aspect, the present application further provides a computer readable storage medium, the computer readable storage medium storing a computer program, when the computer program is executed by a processor, the steps of the above method are performed.

[0014] The present application provides a battery state of health estimation method, device, electronic device and medium, wherein the method comprises: acquiring real-time operation data of a battery, and determining feature data of the battery according to the real-time operation data, the feature data representing an aging trend of the battery; inputting the feature data into a machine learning model to obtain an initial state of health estimation value of the battery, and inputting the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicating a corresponding relationship between historical feature data of the battery and a state of health value of the battery; and inputting the initial state of health estimation value and the ohmic internal resistance real-time identification value into an adaptive Kalman filter for data fusion to obtain a final state of health estimation value of the battery. Through the present application, accurate estimation of the state of health of the battery is realized.

[0015] In order to make the above objectives, characteristics and advantages of the present application more apparent, more comprehensible, the following preferred embodiments are specifically described in detail below, together with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0017] Figure 1 A flow chart of a battery state of health estimation method provided by the embodiments of the present application; Figure 2 A flow chart of constructing a machine learning model provided by the embodiments of the present application; Figure 3 A structural schematic diagram of a battery state of health estimation device provided by the embodiments of the present application; Figure 4 A structural schematic diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to make the objectives, technical solutions and advantages of the embodiments of the present application more apparent, the following will combine the drawings in the embodiments of the present application to make a clear and complete description of the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, every other embodiment obtained by those skilled in the art without creative labor is within the scope of protection of the present application.

[0019] Firstly, the application scenarios applicable to the present application are introduced. The present application can be applied to the field of battery technology.

[0020] Battery state of health (SOH) estimation is one of the core functions of the battery management system, and its accuracy directly affects the reliability and safety of the device operation. At present, although the traditional SOH estimation method has been widely studied, it still has significant limitations.

[0021] The voltage difference method estimates SOH by changes in battery voltage, but has significant drawbacks. First, voltage changes are influenced by factors such as temperature and load, leading to inaccurate SOH estimates. In addition, the voltage change reaction of battery aging is slow, limiting the ability to detect short-term aging. In low power or high load situations, voltage differences may not match the actual SOH, leading to false positives. Moreover, different types of batteries have different sensitivities to voltage changes, reducing the applicability of this method.

[0022] The internal resistance measurement method uses internal resistance changes to assess SOH, but also has many problems. Internal resistance is affected by environmental conditions such as temperature and humidity, resulting in unstable measurement results. Moreover, the change in internal resistance is weak in the early stages of battery aging, making it difficult to reflect the deterioration state of the battery in a timely manner. In addition, accurate measurement of internal resistance requires complex equipment, increasing the cost of the system. External current fluctuations and noise can also interfere with internal resistance measurements, reducing the accuracy of the estimates.

[0023] The charge counter method estimates SOH by counting the number of charge and discharge cycles and accumulating the amount of electricity, but its drawbacks are obvious. First, focusing only on cycle count while ignoring actual performance changes in the battery leads to an inability to accurately reflect the degree of degradation. Second, the method's estimates are not very accurate because it only relies on the number of charge and discharge cycles and cannot provide specific health status information. In addition, irregular charging and discharging processes can affect the accuracy of the count, and the reliance on complete historical data can severely impact the evaluation results if data is missing or recorded incorrectly.

[0024] The model prediction method uses mathematical models to predict SOH, but its limitations are also obvious. First, building a battery model is complex and involves a large number of parameter adjustments, making real-time prediction difficult to achieve. Second, model prediction often requires powerful computing resources, especially in embedded systems, which can increase processing burden. In addition, this method is less adaptable to different types of batteries and cannot be used universally, and must be adjusted for each type of battery. Finally, due to the large amount of computation, real-time prediction can cause delays, especially in high-speed charging and discharging situations.

[0025] Therefore, the embodiments of the present application provide a battery health state estimation method, device, electronic equipment and medium, which aims to accurately estimate the battery health state.

[0026] Please refer to Figure 1 , Figure 1 The flowchart of a battery health state estimation method provided by the embodiments of the present application. As shown in Figure 1 The battery health state estimation method provided by the embodiments of the present application comprises the following steps. S101. Obtain the real-time operating data of the battery and determine the characteristic data of the battery based on the real-time operating data.

[0027] Here, the feature data characterizes the aging trend of the battery.

[0028] In a preferred embodiment of this application, the feature data is determined by: acquiring real-time operating data, which includes the voltage sequence, current sequence, and temperature sequence of the battery in the target charge-discharge cycle; and calculating the feature data based on the voltage sequence, current sequence, and temperature sequence, which includes the slope of the constant current charging voltage plateau and the differential capacity peak of the discharge curve. Specifically, firstly, a complete charge-discharge cycle's voltage, current, and temperature sequence over time is collected. Based on this sequence, the constant-current charging voltage plateau slope and the differential capacity peak of the discharge curve are obtained. The constant-current charging voltage plateau slope is a crucial aging indicator, as the battery ages, its internal chemical active materials decrease, and its internal resistance increases, leading to a faster rate of voltage rise (slope) during constant-current charging. The differential capacity peak of the discharge curve is obtained by differentiating the discharge voltage curve. The peaks on this curve correspond to specific electrochemical reactions within the battery. With aging, the intensity and position of these peaks exhibit a regular shift and decrease. This feature is highly sensitive to aging, and these features reveal the underlying chemical mechanisms of aging earlier and more stably than directly observing capacity, providing high-quality input for machine learning models.

[0029] S102. Input the feature data into the machine learning model to obtain the initial health state estimate of the battery, and input the real-time running data into the equivalent circuit model to obtain the ohmic internal resistance identification value of the battery.

[0030] In a preferred embodiment of this application, the equivalent circuit model is a second-order RC model, expressed by the following formula:

[0031] in, This is the terminal voltage value. Open circuit voltage, For load current, and These are the electrochemical polarization voltage and concentration polarization voltage, respectively, and R is the ohmic internal resistance identification value.

[0032] Here, the machine learning model indicates the correspondence between the historical feature data of the battery and the health state value of the battery, and the machine learning model is preferably a neural network, a support vector machine, etc., which can establish a nonlinear mapping relationship between the battery features and the true SOH through learning. When the current battery feature data is input, the machine learning model will output an initial SOH estimation value according to the learned mapping relationship.

[0033] The specific process of constructing the machine learning model will be introduced below. Figure 2 The specific process of constructing the machine learning model will be introduced below.

[0034] Figure 2 The specific process of constructing the machine learning model will be introduced below.

[0035] S201, obtaining a training sample set of the battery.

[0036] Here, the training sample set includes historical feature data and measured health state estimation values of the battery corresponding to each of a plurality of charge and discharge cycles.

[0037] S202, for each training sample in the training sample set, taking the historical feature data of the training sample as input and taking the measured health state estimation value of the training sample as output, to express the relationship between the feature data and the health state value of the battery in the form of a machine learning model.

[0038] Returning to Figure 1 , S103, inputting the initial health state estimation value and the ohmic internal resistance real-time recognition value into an adaptive Kalman filter for data fusion to obtain the final health state estimation value of the battery.

[0039] Specifically, the initial SOH estimation value may not be accurate under certain working conditions, and therefore it is necessary to fuse the initial SOH estimation value and the ohmic internal resistance recognition value to obtain the final health state estimation value of the battery.

[0040] In a preferred example of the present application, the final health state estimation value of the battery is obtained by inputting the initial health state estimation value and the ohmic internal resistance real-time recognition value into an adaptive Kalman filter and taking them as two observation inputs of the adaptive Kalman filter; according to the recursive estimation algorithm of the adaptive Kalman filter, the credibility of the two observation inputs is dynamically weighted and optimally fused to output the final health state estimation value.

[0041] In a preferred example of the present application, the adaptive Kalman filter is configured in the following manner: the state of health value of the battery is defined as the main state variable, and an observation equation is established to associate the state of health value of the battery with the ohmic internal resistance value; a state transition matrix is preset according to a preset battery aging empirical model, and an observation noise covariance matrix is preset according to a measurement noise level, so as to complete the configuration of the adaptive Kalman filter.

[0042] The objective here is to fuse the two inputs with uncertainties, the initial SOH estimate from machine learning and the ohmic internal resistance identification from equivalent circuit model, into one optimal final state of health estimate. The adaptive Kalman filter first needs to be configured purposefully: the state of health of the battery is explicitly set as the only state variable for accurate estimation; its state equation generally follows the characteristic of slow change of SOH, and is constructed in the form of SOH(k) = SOH(k-1) + process noise, to describe its basic evolution over time; at the same time, a key observation equation is established to establish the mathematical association between the ohmic internal resistance observation value and the state of health. In addition, the state transition and observation noise covariance matrices are preset according to the understanding of the system noise level, to lay the foundation for the dynamic weight allocation of the filter.

[0043] After the configuration of the adaptive Kalman filter is completed, the adaptive Kalman filter enters the workflow of recursive estimation. In each estimation period, it first performs the prediction step, i.e., deduces the SOH prediction value at the current time according to the final SOH estimate value at the last time. Then, in the update step, the filter receives two new observation information - the initial SOH estimate value directly given by machine learning and the ohmic internal resistance value identified by the equivalent circuit model. The adaptive Kalman filter calculates the difference between the prediction value and the two observation values, and evaluates the credibility of each information source in real time according to its internal adaptive algorithm. If the uncertainty of a certain observation source increases at the current time, its weight will be automatically lowered; otherwise, the weight will be increased. Finally, through a set of optimal estimation algorithm, the adaptive Kalman filter fuses its prediction value with the two observation values after weight allocation, and outputs a final state of health estimate value that is statistically closest to the true value. Based on the same inventive concept, the present application also provides a battery state of health estimation device corresponding to the battery state of health estimation method. Since the principle of the device in the present application solves the problem similar to the above-mentioned battery state of health estimation method of the present application, the implementation of the device can be referred to the implementation of the method, and the repeated parts will not be described here.

[0044] Please refer to Figure 3 , Figure 3 The structure of the battery state of health estimation device provided in the present application is shown in the figure. Figure 3The battery health state estimation apparatus 300 includes: The acquisition module 301 is configured to acquire real-time operation data of the battery, and determine feature data of the battery according to the real-time operation data, the feature data representing an aging trend of the battery.

[0045] The prediction module 302 is configured to input the feature data into a machine learning model to obtain an initial health state estimation value of the battery, and input the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicating a corresponding relationship between historical feature data of the battery and a health state value of the battery.

[0046] The correction module 303 is configured to input the initial health state estimation value and the ohmic internal resistance real-time identification value into an adaptive Kalman filter for data fusion to obtain a final health state estimation value of the battery.

[0047] In a preferred example of the present application, the acquisition module is further configured to: acquire the real-time operation data, the real-time operation data including a voltage sequence, a current sequence and a temperature sequence of the battery in a target charge-discharge cycle; and calculate the feature data according to the voltage sequence, the current sequence and the temperature sequence, the feature data including a constant-current charge voltage platform slope and a discharge curve differential capacity peak value.

[0048] Please refer to Figure 4 , Figure 4 A structural schematic diagram of an electronic device provided by an embodiment of the present application is shown in FIG. 4. Figure 4 As shown in FIG. 4, the electronic device 400 includes a processor 410, a memory 420 and a bus 430.

[0049] The memory 420 stores machine-readable instructions executable by the processor 410, and when the electronic device 400 is running, the processor 410 and the memory 420 communicate through the bus 430, and the machine-readable instructions executed by the processor 410 can perform the steps of the battery health state estimation method in the method embodiments shown in Figure 1 and Figure 2 The specific implementation can be referred to the method embodiments, which will not be described here.

[0050] The present application also provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to perform the steps of the battery health state estimation method in the method embodiments shown in Figure 1 and Figure 2 The specific implementation can be referred to the method embodiments, which will not be described here.

[0051] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the system, device and unit described above can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0052] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there can be another division manner in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some communication interfaces, devices or units, and can be electrical, mechanical or other forms.

[0053] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0054] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit.

[0055] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a non-volatile computer readable storage medium executable by a processor. Based on this understanding, the technical solutions of the present application or the essential part or part of the technical solutions that make contributions to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and various storage medium that can store program codes.

[0056] Finally, it should be noted that the above-described embodiments are merely specific embodiments of the present application, which are used to illustrate the technical solutions of the present application, but not to limit the same. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that any skilled person in the art can still modify or easily think of changes to the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some of the technical features, within the technical scope disclosed by the present application. The modifications, changes or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A battery state of health estimation method, characterized by, The method comprises: acquiring real-time operation data of a battery, and determining characteristic data of the battery according to the real-time operation data, the characteristic data representing an aging trend of the battery; inputting the characteristic data into a machine learning model to obtain an initial state of health estimation value of the battery, and inputting the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicating a corresponding relationship between historical characteristic data of the battery and a state of health value of the battery; inputting the initial state of health estimation value and the ohmic internal resistance real-time identification value into an adaptive Kalman filter for data fusion to obtain a final state of health estimation value of the battery.

2. The method of claim 1, wherein, The characteristic data is determined in the following manner: acquiring the real-time operation data, the real-time operation data including a voltage sequence, a current sequence and a temperature sequence of the battery in a target charge-discharge cycle; calculating the characteristic data according to the voltage sequence, the current sequence and the temperature sequence, the characteristic data including a constant-current charge voltage platform slope and a discharge curve differential capacity peak value.

3. The method of claim 2, wherein, The machine learning model is constructed in the following manner: acquiring a training sample set of the battery, the training sample set including respective historical characteristic data and measured state of health estimation values of the battery in multiple charge-discharge cycles; for each training sample in the training sample set, inputting the historical characteristic data of the training sample as input and inputting the measured state of health estimation value of the training sample as output, so as to represent a relationship between the characteristic data and the state of health value of the battery by the machine learning model.

4. The method of claim 3, wherein, The final state of health estimation value of the battery is obtained in the following manner: inputting the initial state of health estimation value and the ohmic internal resistance real-time identification value into an adaptive Kalman filter as two observation inputs of the adaptive Kalman filter; dynamically weighing the credibility of the two observation inputs and performing data fusion according to a recursive estimation algorithm of the adaptive Kalman filter to output the final state of health estimation value.

5. The method of claim 4, wherein, The adaptive Kalman filter is configured in the following manner: defining the state of health value of the battery as a main state variable, and establishing an observation equation relating the state of health value of the battery to an ohmic internal resistance value; presetting a state transition matrix according to a preset battery aging empirical model, and presetting an observation noise covariance matrix according to a measurement noise level, so as to complete the configuration of the adaptive Kalman filter.

6. The method of claim 1, wherein, The equivalent circuit model is represented by the following formula: wherein, V is the terminal voltage value, OCV is the open circuit voltage value, I is the current value, and U is the electrochemical polarization voltage value and U is the concentration polarization voltage value, respectively, and R is the ohmic internal resistance recognition value.

7. A battery state of health estimation device characterized by comprising: The device comprises: an acquisition module configured to acquire real-time operation data of a battery, and determine characteristic data of the battery according to the real-time operation data, the characteristic data representing an aging trend of the battery; a prediction module configured to input the characteristic data into a machine learning model to obtain an initial state of health estimation value of the battery, and input the real-time operation data into an equivalent circuit model to obtain an ohmic internal resistance identification value of the battery, the machine learning model indicating a corresponding relationship between historical characteristic data of the battery and a state of health value of the battery; and an adaptive Kalman filter configured to input the initial state of health estimation value and the ohmic internal resistance real-time identification value into the adaptive Kalman filter for data fusion to obtain a final state of health estimation value of the battery. A correction module is configured to input the initial health state estimation value and the real-time ohmic internal resistance recognition value into an adaptive Kalman filter for data fusion to obtain a final health state estimation value of the battery.

8. The apparatus of claim 7, wherein, The acquisition module is further configured to: acquire the real-time operation data, the real-time operation data including a voltage sequence, a current sequence and a temperature sequence of the battery in a target charge-discharge cycle; calculate the feature data according to the voltage sequence, the current sequence and the temperature sequence, the feature data including a constant-current charge voltage platform slope and a discharge curve differential capacity peak value.

9. An electronic device, comprising: comprise: a processor, a memory and a bus, the memory storing machine readable instructions executable by the processor, the processor and the memory communicating via the bus when the electronic device is running, the processor executing the machine readable instructions to perform the steps of the method of any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to perform the steps of the method of any one of claims 1 to 6.

Images