Method for detecting residual electric quantity of lithium ion battery

Abstract

The invention relates to the technical field of lithium batteries, and particularly discloses a lithium ion battery remaining capacity detection method, which comprises the following steps: firstly, carrying out preliminary SOC calculation based on an ampere-hour integral method to obtain a preliminary SOC predicted value of a battery, and estimating dynamic internal resistance of the battery by combining the preliminary SOC predicted value of the battery and related electrical parameters to reflect the internal state of the battery; and furthermore, time sequence analysis and joint reasoning are performed on the real-time internal resistance data and the real-time surface temperature data of the battery so as to estimate a battery capacity attenuation factor, and real-time correction is performed on the actual capacity of the battery based on the battery capacity attenuation factor, so that a more accurate estimated value of the remaining capacity of the battery is obtained. Through the mode, the influence of battery aging on battery remaining capacity detection can be effectively compensated, and the SOC detection accuracy and reliability of the lithium ion battery in the whole life cycle in various application scenes such as electric automobiles and portable electronic equipment can be improved.

Description

Technical Field

[0001] The present application relates to the field of lithium battery technology, and more specifically, to a method for detecting the remaining power of a lithium-ion battery. Background Art

[0002] With the rapid development of energy technology, lithium-ion batteries have been widely used in electric vehicles, portable electronic devices, energy storage systems, and other fields due to their high energy density, long cycle life, and low self-discharge rate. In this context, accurate and real-time knowledge of the remaining capacity (State of Charge, SOC) of lithium-ion batteries is crucial for ensuring safe device operation, optimizing energy management strategies, and improving user experience. For example, in electric vehicles, accurate SOC estimation can effectively alleviate users' range anxiety and provide critical data to the battery management system (BMS) to prevent battery overcharging or over-discharging, thereby extending battery life and ensuring driving safety.

[0003] At present, the estimation methods of lithium-ion battery SOC mainly include the open circuit voltage method, the ampere-hour integration method (coulomb counting method), etc. Although the open circuit voltage method is simple, it requires the battery to be stationary for a long time to achieve a stable open circuit voltage, which makes it difficult to meet the real-time detection requirements under dynamic working conditions. The ampere-hour integration method calculates the SOC by accumulating the amount of charge flowing into or out of the battery. It is widely used because of its simple principle and strong real-time performance. However, this method relies on fixed battery nominal capacity parameters and cannot dynamically adapt to the actual capacity fluctuations caused by factors such as aging, attenuation, and temperature changes in actual use. Long-term use will cause error accumulation, causing the SOC estimation result to gradually deviate from the true value. Although some solutions attempt to improve accuracy by regularly calibrating capacity parameters, it is often difficult to fundamentally solve the error accumulation problem because the calibration frequency is difficult to match the dynamic nature of the actual battery capacity changes.

[0004] Therefore, an optimized method for detecting the remaining capacity of a lithium-ion battery is desired. Summary of the Invention

[0005] In order to solve the above technical problems, the present application is proposed. An embodiment of the present application provides a method for detecting the remaining capacity of a lithium-ion battery, which first performs a preliminary SOC calculation based on the ampere-hour integration method to obtain a preliminary SOC prediction value of the battery, and estimates the dynamic internal resistance of the battery in combination with the preliminary SOC prediction value of the battery and related electrical parameters to reflect the internal state of the battery; then, the battery capacity attenuation factor is estimated by performing time series analysis and joint reasoning on the real-time internal resistance data and real-time surface temperature data of the battery, and the actual capacity of the battery is corrected in real time based on the battery capacity attenuation factor, thereby obtaining a more accurate estimate of the remaining capacity of the battery. In this way, the influence of battery aging on the detection of the remaining capacity of the battery can be effectively compensated, and the accuracy and reliability of the SOC detection of lithium-ion batteries in various application scenarios such as electric vehicles and portable electronic devices throughout their life cycle can be improved.

[0006] According to one aspect of the present application, a method for detecting the remaining power of a lithium-ion battery is provided, comprising:

[0007] Collect real-time current data, real-time terminal voltage data and real-time surface temperature data of lithium-ion batteries within a preset time window;

[0008] Calculating preliminary SOC real-time prediction data based on the real-time current data and the actual battery capacity value of the previous cycle;

[0009] Calculating the real-time internal resistance data of the battery within the preset time window based on the preliminary SOC real-time prediction data;

[0010] Inputting the real-time internal resistance data and the real-time surface temperature data of the battery into a battery capacity attenuation evaluation module to obtain a battery capacity attenuation factor;

[0011] Based on the battery capacity attenuation factor, the battery actual capacity value of the previous cycle is updated to obtain an updated battery actual capacity value, and based on the updated battery actual capacity value, the preliminary SOC real-time prediction data is corrected to obtain battery SOC real-time detection data.

[0012] Compared with the prior art, the lithium-ion battery remaining capacity detection method provided by this application first performs a preliminary SOC calculation based on the ampere-hour integration method to obtain the preliminary SOC prediction value of the battery, and estimates the dynamic internal resistance of the battery in combination with the preliminary SOC prediction value of the battery and related electrical parameters to reflect the internal state of the battery; then, by performing time series analysis and joint reasoning on the real-time internal resistance data and real-time surface temperature data of the battery, the battery capacity attenuation factor is estimated, and the actual battery capacity is corrected in real time based on the battery capacity attenuation factor, thereby obtaining a more accurate estimate of the remaining battery capacity. In this way, the impact of battery aging on the detection of the remaining battery capacity can be effectively compensated, and the accuracy and reliability of the SOC detection of lithium-ion batteries in various application scenarios such as electric vehicles and portable electronic devices throughout their life cycle can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0014] Figure 1 Flowchart of a method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application.

[0015] Figure 2 Schematic diagram of data flow of a method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application.

[0016] Figure 3 4 is a flowchart of sub-step S4 of the method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application.

[0017] Figure 4 4 is a flowchart of sub-step S42 of the method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] As used in this application and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not intended to refer to the singular but may include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0019] Although the present application makes various references to certain modules in the system according to embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are illustrative only, and different aspects of the system and method can use different modules.

[0020] Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the various steps may be processed in reverse order or simultaneously, as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0022] It is worth noting that in this application, all actions to obtain data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0023] In response to the technical problems described in the above background technology, this application proposes a method for detecting the remaining capacity of a lithium-ion battery, which first performs a preliminary SOC calculation based on the ampere-hour integration method to obtain a preliminary SOC prediction value of the battery, and estimates the dynamic internal resistance of the battery in combination with the preliminary SOC prediction value of the battery and related electrical parameters to reflect the internal state of the battery; then, by performing time series analysis and joint reasoning on the real-time internal resistance data and real-time surface temperature data of the battery, the battery capacity attenuation factor is estimated, and the actual battery capacity is corrected in real time based on the battery capacity attenuation factor, thereby obtaining a more accurate estimate of the remaining capacity of the battery. In this way, the effect of battery aging on the detection of the remaining capacity of the battery can be effectively compensated, and the accuracy and reliability of the SOC detection of lithium-ion batteries in various application scenarios such as electric vehicles and portable electronic devices throughout their life cycle can be improved.

[0024] Figure 1 Flowchart of a method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application. Figure 1As shown, the method for detecting the remaining capacity of a lithium-ion battery includes the following steps: S1, collecting real-time current data, real-time terminal voltage data and real-time surface temperature data of the lithium-ion battery within a preset time window; S2, calculating preliminary SOC real-time prediction data based on the real-time current data and the actual capacity value of the battery in the previous cycle; S3, calculating the real-time internal resistance data of the battery in the preset time window based on the preliminary SOC real-time prediction data; S4, inputting the real-time internal resistance data and the real-time surface temperature data of the battery into a battery capacity attenuation evaluation module to obtain a battery capacity attenuation factor; S5, updating the actual capacity value of the battery in the previous cycle based on the battery capacity attenuation factor to obtain an updated actual battery capacity value, and correcting the preliminary SOC real-time prediction data based on the updated actual battery capacity value to obtain battery SOC real-time detection data.

[0025] In the above-mentioned method for detecting the remaining capacity of a lithium-ion battery, step S1 collects the real-time current data, real-time terminal voltage data, and real-time surface temperature data of the lithium-ion battery within a preset time window. It should be understood that during the charging and discharging process of a lithium-ion battery, the dynamic changes in current and voltage are closely related to the remaining capacity (SOC), and the actual capacity of the battery is significantly affected by temperature. Therefore, in order to obtain complete battery status information to support subsequent calculation of the remaining battery capacity, the present application uses sensors to monitor and record the current, terminal voltage, and surface temperature data within a preset time window during battery operation in real time to achieve a comprehensive perception of the battery status.

[0026] In the specific implementation process, the data acquisition system is usually composed of multiple sensors, which are used to monitor key parameters such as battery current, voltage and temperature. Among them, the current sensor is used to measure the instantaneous current value during the battery charging and discharging process. Its installation position is generally located in the main circuit of the battery so that it can accurately capture the amount of charge flowing through the battery. Since the current variation range of lithium-ion batteries under different operating conditions is large, the selected current sensor is required to have a wide range, high resolution and fast response characteristics to meet the real-time monitoring needs under dynamic working conditions. In addition, to improve measurement accuracy and reduce external interference, the current sensor should adopt a high-precision Hall effect or shunt resistor structure and be equipped with a signal conditioning circuit to ensure the stability and consistency of the output signal.

[0027] Meanwhile, the voltage sensor collects terminal voltage data across the battery. Terminal voltage is a key indicator of the battery's internal chemical reaction state and has a nonlinear relationship with the SOC. Especially during dynamic charge and discharge, the changing trend of the terminal voltage provides a crucial reference for subsequent algorithms. Therefore, the voltage sensor should be installed as close as possible to the positive and negative battery terminals to avoid measurement errors caused by line voltage drops. Furthermore, given the potential imbalance among individual cells in a battery pack, the voltage acquisition module should feature multi-channel simultaneous sampling to ensure independent monitoring of the terminal voltage of each cell. Furthermore, the voltage sensor should have a high input impedance to minimize the impact on the battery's internal operating state, and be equipped with appropriate filtering and protection circuits to prevent data distortion due to overvoltage or noise interference.

[0028] In addition to current and voltage parameters, temperature is also a significant factor influencing the capacity and SOC estimation accuracy of lithium-ion batteries. During battery operation, internal chemical reactions and ohmic losses generate a certain amount of heat, causing surface temperature fluctuations. These temperature fluctuations not only affect the conductivity of the electrolyte and the activity of the electrode materials, but also further impact the battery's actual usable capacity. Therefore, this solution incorporates a real-time monitoring mechanism for battery surface temperature, using temperature sensors to collect temperature data from the battery casing or electrode connections. These temperature sensors, typically NTC thermistors or digital temperature sensors, are installed in key hot spots on the battery surface and communicate with the BMS system for continuous data collection and transmission. To ensure representative temperature data, sensor placement should comprehensively consider the battery's structural layout, heat dissipation paths, and possible local hot spots. If necessary, a multi-point distribution approach can be used to obtain more comprehensive temperature field information.

[0029] In order to ensure the time synchronization and spatial consistency of the collected data, the entire data acquisition system must also be equipped with a unified clock source and a high-speed data processing unit. All raw data collected by the sensors must undergo analog-to-digital conversion, be preliminarily processed by the central processing unit, and be segmented and stored according to the set time window. The preset time window refers to a fixed time period set according to the operating frequency of the battery and the system response requirements, such as 1 second, 5 seconds or 10 seconds. Within this time window, the system will continuously collect and record current, voltage and temperature data to form a complete data sequence for subsequent analysis and modeling. The selection of the time window should take into account the balance between real-time performance and data integrity. If the window is too short, it may lead to large data fluctuations and difficulty in feature extraction; if the window is too long, it will reduce the response speed of the system and affect the timeliness of SOC estimation.

[0030] Furthermore, during the data acquisition process, environmental factors must be effectively controlled to reduce the impact of external interference on measurement results. For example, in electric vehicles or energy storage systems, batteries are often located in complex electromagnetic environments and are susceptible to interference from high-voltage equipment such as motor controllers and DC-DC converters. Therefore, when designing the acquisition system, shielding, grounding, and isolation measures should be implemented to ensure stable signal transmission. Furthermore, for different application scenarios, such as high temperature, low temperature, or high humidity conditions, sensor selection and packaging methods must be optimized to ensure long-term reliable operation in harsh environments.

[0031] It is worth noting that during the actual operation of lithium-ion batteries, the changes in current, voltage, and temperature do not exist in isolation, but are interrelated and work together. For example, under high-current discharge conditions, the internal polarization effect of the battery is enhanced, causing a significant drop in terminal voltage; while under low-temperature conditions, the electrolyte viscosity increases and the ion diffusion rate decreases, which can also cause the voltage platform to shift downward and the internal resistance to increase. These complex phenomena will directly affect the accuracy of SOC calculations. Therefore, relying solely on a single parameter is difficult to fully reflect the state of the battery. Only through the coordinated collection and joint analysis of multiple parameters can more abundant input information be provided for the subsequent SOC estimation model.

[0032] In the above-mentioned method for detecting the remaining capacity of a lithium-ion battery, the step S2 calculates preliminary SOC real-time prediction data based on the real-time current data and the actual capacity value of the battery in the previous cycle. It should be understood that since the ampere-hour integration method has the characteristics of simple principle and strong real-time performance, it is suitable for preliminary SOC estimation in dynamic scenarios. Therefore, in order to quickly obtain the real-time change trend of SOC under dynamic working conditions, the present application is based on the ampere-hour integration technology, and uses the actual capacity value updated in the previous cycle to integrate the current of the current cycle, and calculates the preliminary SOC prediction value at each time point by point-by-point recursion, so as to achieve real-time tracking of battery power consumption or replenishment. Here, the actual capacity value of the battery in the previous cycle is calculated and updated based on the battery remaining capacity detection method described in this application within the previous preset time window of the current preset time window, and its specific calculation process is described in detail in the subsequent specification.

[0033] Specifically, the preliminary SOC real-time prediction value at each time point is calculated using the following SOC calculation formula, wherein the SOC calculation formula is as follows:

[0034]

[0035] Where Δt represents the sampling interval, I avg represents the average current value within the sampling interval, k represents the current sampling time point, C represents the actual capacity value of the battery, and SOC(·) represents the preliminary SOC real-time prediction value at the corresponding sampling time point.

[0036] In the above-mentioned method for detecting the remaining capacity of a lithium-ion battery, the step S3 calculates the real-time internal resistance data of the battery within the preset time window based on the preliminary SOC real-time prediction data. It should be understood that the internal resistance of a battery is a key parameter reflecting its health status, charge and discharge capabilities, and energy conversion efficiency, and is closely related to the battery's SOC, temperature, and aging degree. Its real-time changes can provide important clues for the evaluation of battery performance attenuation. Therefore, in order to obtain dynamic internal resistance parameters that can characterize the current performance status and aging signs of the battery, and provide key inputs for subsequent capacity attenuation evaluations, this application is based on Ohm's law and the principle of the battery equivalent circuit model, and utilizes the battery open circuit voltage data corresponding to the preliminary SOC prediction value at each time point (according to a pre-calibrated SOC-V OCV The instantaneous value of the battery's internal resistance at each time point is calculated using real-time terminal voltage and current data (using the curve acquisition). By establishing a connection between externally measurable parameters and internal states, the fluctuation of real-time internal resistance can dynamically reflect the changes in the battery's polarization degree and aging trend during the charge and discharge process, providing dynamic feature input for subsequent battery capacity decay assessment.

[0037] In the specific implementation process, first, according to SOC-V OCV The battery open circuit voltage data is obtained from the curve; then, the real-time internal resistance value of the battery at each time point is calculated using the following battery internal resistance calculation formula, wherein the battery internal resistance calculation formula is:

[0038]

[0039] Among them, V OCV Indicates the battery open circuit voltage value, V 端 represents the real-time terminal voltage value, I represents the real-time current value, |·| represents the absolute value, R in Indicates the real-time internal resistance of the battery.

[0040] In the above-mentioned method for detecting the remaining power of a lithium-ion battery, in step S4, the real-time internal resistance data of the battery and the real-time surface temperature data are input into the battery capacity attenuation evaluation module to obtain the battery capacity attenuation factor. Specifically, considering that the battery capacity attenuation is the combined result of the internal resistance increase and the temperature change, a single internal resistance or temperature parameter cannot accurately characterize the degree of attenuation. Therefore, in order to dynamically evaluate the degree of battery capacity attenuation, the present application further performs time series analysis and joint reasoning on the real-time internal resistance data of the battery and the real-time surface temperature data to comprehensively consider the influence of the internal resistance change and the temperature effect on the battery capacity attenuation, thereby obtaining the battery capacity attenuation factor. Among them, Figure 3 FIG. 4 is a flow chart of sub-step S4 of the method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application. Figure 3As shown, the step S4 includes the following steps: S41, respectively extracting the time series fluctuation characteristics of the battery real-time internal resistance data and the real-time surface temperature data to obtain a battery internal resistance time series fluctuation characteristic coding vector and a battery surface temperature time series fluctuation characteristic coding vector; S42, performing time series interactive response reasoning on the battery internal resistance time series fluctuation characteristic coding vector and the battery surface temperature time series fluctuation characteristic coding vector to obtain a battery capacity attenuation state characteristic coding vector; S43, performing feature decoding on the battery capacity attenuation state characteristic coding vector to obtain the battery capacity attenuation factor.

[0041] Specifically, in a specific example of the present application, the step S41 includes: performing time series fluctuation feature extraction based on one-dimensional convolution coding on the real-time internal resistance data of the battery and the real-time surface temperature data to obtain the battery internal resistance time series fluctuation feature coding vector and the battery surface temperature time series fluctuation feature coding vector. It should be understood that the internal resistance and temperature data of the battery have dynamic time series characteristics. The time series changes of the battery internal resistance reflect the dynamic evolution of the internal structure and chemical characteristics of the battery, and the time series fluctuations of the surface temperature are closely related to the thermal management efficiency of the battery and the external environmental conditions. However, due to the high dimension and high noise of the original data, it is difficult to use it directly for parameter estimation. Therefore, in order to extract effective features from complex time series data and reduce the data dimension, the present application is based on the one-dimensional convolutional neural network (1D-CNN) in deep learning. By using a one-dimensional sliding convolution kernel, the real-time internal resistance data of the battery and the real-time surface temperature data are respectively subjected to local feature encoding to capture the local correlation and time series fluctuation pattern in the internal resistance data and the surface temperature data. Specifically, the internal resistance sequence (such as the resistance values ​​of the last 100 sampling points) and the temperature sequence (such as the temperature values ​​at the corresponding time) within the preset time window are respectively input into the 1D-CNN model. The model adopts a multi-layer 1D-CNN structure, and each layer contains convolution kernels of different scales (such as 3-point, 5-point, and 7-point windows) to extract features of the input time series data layer by layer. After each layer of convolution, the ReLU activation function is used to introduce nonlinearity, and then the feature dimension is reduced through the pooling layer (such as maximum pooling) to gradually abstract the impedance transition characteristics of the internal resistance data and the heat dissipation / heating trend characteristics of the temperature data, and finally output the fixed-dimensional battery internal resistance time series fluctuation feature encoding vector and the battery surface temperature time series fluctuation feature encoding vector. In this way, the noise interference in the real-time internal resistance data and the real-time surface temperature data of the battery can be effectively filtered out and the key time series patterns can be retained, providing high-information-density input features for subsequent joint interactive reasoning.

[0042] Specifically, the step S42 performs time-series interactive response reasoning on the battery internal resistance time-series fluctuation feature coding vector and the battery surface temperature time-series fluctuation feature coding vector to obtain a battery capacity decay state feature coding vector. It should be understood that since battery capacity decay is the result of the combined effect of internal resistance growth and temperature effect, for example, a high temperature environment will accelerate the rise of internal resistance and aggravate the loss of active substances. There is a complex coupling relationship between the two, and traditional simple feature fusion methods such as cascade splicing cannot effectively capture the nonlinear interaction between the two for the characterization of battery capacity decay. Based on this, the present application proposes an interactive response reasoning method, which deconstructs the global time-series interactive relationship between the battery internal resistance and surface temperature into a series of ordered local time-series interactive response fragments, and captures the dynamic evolution and contextual dependency between each local time-series interactive response fragment based on the idea of ​​sequence modeling, so as to achieve in-depth mining of the battery capacity decay state characteristics and obtain a battery capacity decay state feature coding vector. Among them, Figure 4 FIG. 4 is a flow chart of sub-step S42 of the method for detecting the remaining power of a lithium-ion battery according to an embodiment of the present application. Figure 4 As shown, the step S42 includes the steps of: S421, calculating the full-view microscale characteristic response matrix of the battery surface temperature time-series fluctuation characteristic coding vector relative to the battery internal resistance time-series fluctuation characteristic coding vector to obtain the full-view microscale characteristic response matrix of the battery surface temperature-internal resistance time-series fluctuation characteristic; S422, performing multi-dimensional feature screening on the full-view microscale characteristic response matrix of the battery surface temperature-internal resistance time-series fluctuation characteristic to obtain the sequence distribution of the battery surface temperature-internal resistance time-series fluctuation characteristic local response coding vector; S423, performing global coupling correlation deduction on the sequence distribution of the battery surface temperature-internal resistance time-series fluctuation characteristic local response coding vector to obtain the battery capacity decay state characteristic coding vector.

[0043] More specifically, step S421 is expressed as follows:

[0044]

[0045] Where tanh(·) represents the hyperbolic tangent function, v1 represents the characteristic encoding vector of the battery surface temperature time series fluctuation, v2 represents the characteristic encoding vector of the battery internal resistance time series fluctuation, (·) T represents the transpose of the vector, Sigmoid(·) represents the sigmoid activation function, W α is the dynamic weight matrix, W β is the projection matrix, ⊙ represents the point multiplication by position, It means adding by position. represents matrix multiplication, is the hyperbolic tangent asymmetric interaction, is the element-level gated interaction, and M represents the full-view microscale feature response matrix of the battery surface temperature-internal resistance timing fluctuation characteristics.

[0046] That is, by calculating the potential interaction and correlation strength between the battery surface temperature time-series fluctuation feature encoding vector and the battery internal resistance time-series fluctuation feature encoding vector, the interactive relationship between the two in the time-series dimension can be comprehensively and meticulously characterized, and an interactive map that can comprehensively reflect the subtle semantic differences, resonance and mutual influence between the battery surface temperature and internal resistance in the process of time-series fluctuation is generated, that is, the full-view micro-scale feature response matrix of the battery surface temperature-internal resistance time-series fluctuation feature, thereby providing input with high information density and rich interaction details for subsequent capacity decay state inference, laying a key foundation for real-time correction of the actual capacity of the battery and improving the accuracy of lithium-ion battery remaining power detection.

[0047] In particular, when calculating the full-view microscale characteristic response matrix of the battery surface temperature-internal resistance time series fluctuation characteristics, due to the fine-grained correlation matrix The dynamic weight matrix W is used α and the projection matrix W β Modulation is performed in different directions, and the functions tanh and sigmoid have different power orders of exponential functions based on natural constants. This ensures the completeness of the global topological interaction while inducing differentiated dynamic coupling channels for the implicit semantic representation at fine-grained level, thereby affecting the fusion effect of the two parts. Therefore, in order to achieve the equivalence of coupling channels, it is expected to ensure the topological irreducibility of the topological feature manifold based on the heterogeneous coupling channel, that is, to improve the robustness constraint domain of the topological feature manifold for different interaction paths, thereby improving the multimodal fusion gain of the implicit semantic encoding at fine-grained level. Based on this, in a preferred example of the present application, the step S422 includes: first, performing a dual-path ground state interaction degenerate optimization on the full-view microscale feature response matrix of the battery surface temperature-internal resistance timing fluctuation feature to obtain an optimized full-view microscale feature response matrix of the battery surface temperature-internal resistance timing fluctuation feature.

[0048] Specifically, first let:

[0049]

[0050] Among them, M1 represents the battery surface temperature-internal resistance characteristic hyperbolic tangent asymmetric interaction matrix, and M2 represents the battery surface temperature-internal resistance characteristic element-level gated interaction matrix.

[0051] Secondly, the dual-path interaction transition matrix is ​​obtained through the projection invariance order interaction verification, which satisfies:

[0052] M1M3=M3M2

[0053] Wherein, M3 represents the dual-path interaction transition matrix.

[0054] That is, it is considered that different projection parts substantially have an interactive relationship with each other, so that the dual-path interaction transition matrix M3 has corresponding algebraic generators and inverse element mappers based on different interaction orders.

[0055] Then, the ground state decoherence degree of the dual-path interaction transition matrix M3 can be controlled, that is:

[0056]

[0057] in,(·) -1 represents the inverse matrix, ||·|| F represents the F norm, and M4 represents the correction matrix.

[0058] That is, the dual-path interaction transition matrix M3 and its inverse matrix M3 -1 The group representations are used as the algebraic generating kernel and the inverse element mapper respectively to simplify the distribution state through the F norm of the matrix, that is, the group theory distribution state representation is irreducibly degenerated to obtain the ground state coupling topological mapping, so that the directed evolution characteristics of the algebraic generating kernel and the inverse element mapper can be realized by differential encoding of the ground state transition.

[0059] Finally, by correcting M1 and M2 respectively through the correction matrix M4 and then fusing them, the fusion expression effect of the full-view micro-scale feature response matrix of the battery surface temperature-internal resistance time series fluctuation characteristics can be improved, that is:

[0060] M′1=M4M1

[0061] M′2=M4M2

[0062]

[0063] Among them, M'1 represents the hyperbolic tangent asymmetric interaction matrix of the corrected battery surface temperature-internal resistance characteristics, M'2 represents the element-level gated interaction matrix of the corrected battery surface temperature-internal resistance characteristics, and M' represents the full-view microscale characteristic response matrix of the optimized battery surface temperature-internal resistance timing fluctuation characteristics.

[0064] Then, the optimized battery surface temperature-internal resistance time series fluctuation feature full-view micro-scale feature response matrix is ​​subjected to multi-dimensional feature screening according to column vectors to obtain the sequence distribution of the battery surface temperature-internal resistance time series fluctuation feature local response encoding vector, which is expressed as follows:

[0065] Decompose(M')={c1,c2,...,c i ,...,c t}

[0066] Among them, Decompose(·) represents the multidimensional feature screening function, c1, c2, c i and c t They respectively represent the 1st, 2nd, i-th and t-th local response coding vectors of the battery surface temperature-internal resistance time-series fluctuation characteristics in the sequence distribution of the local response coding vectors.

[0067] That is, from the global interaction information of the full-view micro-scale feature response matrix of the optimized battery surface temperature-internal resistance timing fluctuation characteristics, the local feature response with temporal dynamic correlation is extracted, and the global interaction relationship carried by the two-dimensional matrix is ​​converted into a one-dimensional sequence distribution by decomposing it according to the column vector, so as to generate the sequence distribution of the local response encoding vector of the battery surface temperature-internal resistance timing fluctuation characteristics, and form a local information flow that can simulate the gradual reasoning process of human thinking, so that each battery surface temperature-internal resistance timing fluctuation characteristic local response encoding vector corresponds to the local correlation information of the temperature and internal resistance fluctuation characteristics in a specific timing stage or interactive perspective. Its sequence distribution completely retains the temporal interaction logic and dynamic evolution law implicit in the full-view micro-scale feature response matrix of the optimized battery surface temperature-internal resistance timing fluctuation characteristics, so that the subsequent analysis of the battery capacity decay state can be carried out based on the feature sequence with temporal continuity and local focus, thereby more accurately capturing the differentiated effects of temperature and internal resistance fluctuations on battery capacity decay at different stages.

[0068] More specifically, in a specific example of the present application, step S423 includes: inputting the sequence distribution of the local response encoding vector of the battery surface temperature-internal resistance time series fluctuation characteristic into the feature response deduction architecture based on the forward LSTM model to obtain the battery capacity attenuation state feature encoding vector, which is expressed as follows:

[0069] v f =LSTM({c1,c2,...,c i ,...,c t})

[0070] Among them, LSTM(·) represents the forward LSTM model, v f Represents the feature encoding vector of the battery capacity attenuation state.

[0071] That is, by utilizing the unique time series information processing capability of the forward LSTM model, the sequence distribution of the local response encoding vector of the battery surface temperature-internal resistance time series fluctuation characteristics is dynamically modeled. By integrating the local interaction information of each time series stage and its historical context dependency, a battery capacity decay state feature encoding vector that can comprehensively characterize the battery capacity decay state is generated, effectively capturing the long-term dependency and dynamic evolution logic of the temperature and internal resistance fluctuation characteristics at different time steps, so that the output battery capacity decay state feature encoding vector not only contains the independent semantic information of each local interaction, but also deeply encodes the synergistic influence pattern and attenuation feature evolution law of the temperature and internal resistance fluctuations in the time series dimension, providing high-order feature input that integrates time series dynamic correlation for accurate estimation of the battery capacity decay factor, helping to achieve real-time and accurate correction of the actual battery capacity.

[0072] Specifically, in step S43, the battery capacity decay state feature encoding vector is feature decoded to obtain the battery capacity decay factor. That is, in order to map the high-dimensional abstract feature representation of the battery capacity decay state into an interpretable scalar output, this application is based on the regression principle of machine learning, and a decoder is constructed through a fully connected layer and a Sigmoid activation function to perform feature decoding on the battery capacity decay state feature encoding vector to convert it into a specific capacity decay factor. Specifically, first, the battery capacity decay state feature encoding vector is input into a two-layer fully connected network. The first layer performs feature dimensionality reduction through linear transformation and adopts the LeakyReLU activation function to further extract high-order features; the second layer maps the high-order features to a one-dimensional output space through linear transformation, and uses the Sigmoid activation function to compress the output value to between 0 and 1 to obtain the final battery capacity decay factor. In addition, the mean square error (MSE) loss function is used in the training phase of the decoder, and the network weight parameters are optimized through the back propagation algorithm to minimize the error between the predicted value and the true value. In this way, end-to-end decay factor prediction can be achieved, providing an important basis for battery health management and remaining power detection.

[0073] In the above-mentioned method for detecting the remaining capacity of a lithium-ion battery, in step S5, based on the battery capacity attenuation factor, the actual capacity value of the battery in the previous cycle is updated to obtain an updated actual capacity value of the battery, and based on the updated actual capacity value of the battery, the preliminary SOC real-time prediction data is corrected to obtain battery SOC real-time detection data. It should be understood that since the core error source of the traditional ampere-hour integration method is the deviation between the rated capacity parameter and the actual capacity, and the capacity attenuation factor can quantify the degree of degradation of the current state of the battery relative to the rated capacity. Therefore, in order to achieve online self-correction of the battery capacity parameters, the present application dynamically adjusts the battery rated capacity parameters based on the battery capacity attenuation factor to obtain an updated actual capacity value of the battery in the current cycle. Specifically, the initial battery rated capacity is used as a reference and multiplied by the battery capacity attenuation factor to obtain the actual capacity value of the battery in the current cycle. Furthermore, since the battery capacity parameter used in the ampere-hour integration method is the actual battery capacity value of the previous cycle, after obtaining the updated actual battery capacity value of the current cycle, this application replaces the actual battery capacity value of the previous cycle with it and re-executes the ampere-hour integration process based on the updated actual battery capacity value to correct the integration error caused by capacity deviation, thereby obtaining more accurate real-time battery SOC detection data. In this way, dynamic and accurate monitoring of battery SOC can be achieved, effectively improving the accuracy and reliability of the battery management system.

[0074] Furthermore, in the above-mentioned method for detecting the remaining capacity of a lithium-ion battery, it also includes: in response to the real-time terminal voltage value of the lithium-ion battery reaching the charge cut-off voltage or the discharge cut-off voltage, resetting the battery SOC real-time detection data to 100% or 0%. Specifically, since during the charging and discharging process of the lithium-ion battery, when the real-time terminal voltage reaches the charge cut-off voltage or the discharge cut-off voltage, it indicates that the battery is in a fully charged or fully discharged state, the remaining capacity of the battery at this time has a clear physical boundary (100% for full charge and 0% for full discharge), and the traditional ampere-hour integration method may cause the SOC estimation value at this time to deviate from the true value due to error accumulation. Therefore, in order to realize the forced calibration of SOC at the critical point where the battery state is clear, eliminate the influence of long-term error accumulation and ensure that the detection results are consistent with physical reality, this application sets a voltage-triggered SOC reset mechanism based on the battery safe working boundary principle. Specifically, the terminal voltage data is monitored in real time in the battery management system (BMS). When the terminal voltage is detected to reach the preset charge cut-off voltage (such as 4.2V), no matter what the current SOC real-time detection data is, it is forced to reset to 100%; when the terminal voltage drops to the discharge cut-off voltage (such as 2.75V), it is reset to 0%. This process can be implemented through conditional judgment statements in hardware circuits or software programs. For example, a voltage threshold comparator is set in the control logic of the microcontroller (MCU), and when the voltage signal exceeds the threshold, the interrupt program is triggered to perform a reset operation. In this way, the hard threshold of the battery terminal voltage is used as the benchmark for SOC calibration, and the accumulated errors caused by capacity parameter deviation, integral error, etc. are directly corrected, which can effectively avoid the risk of "overcharging" or "overdischarging".

[0075] In summary, a method for detecting the remaining capacity of a lithium-ion battery based on an embodiment of the present application is explained. It first performs a preliminary SOC calculation based on the ampere-hour integration method to obtain a preliminary SOC prediction value of the battery, and estimates the dynamic internal resistance of the battery in combination with the preliminary SOC prediction value of the battery and related electrical parameters to reflect the internal state of the battery; furthermore, the battery capacity attenuation factor is estimated by performing time series analysis and joint reasoning on the real-time internal resistance data and real-time surface temperature data of the battery, and the actual capacity of the battery is corrected in real time based on the battery capacity attenuation factor, thereby obtaining a more accurate estimate of the remaining capacity of the battery. In this way, the effect of battery aging on the detection of the remaining capacity of the battery can be effectively compensated, and the accuracy and reliability of the SOC detection of lithium-ion batteries in various application scenarios such as electric vehicles and portable electronic devices throughout their life cycle can be improved.

[0076] The basic principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in the present invention are merely illustrative and non-limiting, and should not be construed as necessarily possessed by each embodiment of the present invention. Furthermore, the specific details of the above embodiments are provided for illustrative purposes and to facilitate understanding, and are not intended to be limiting. These details do not necessarily limit the present invention to being implemented using these specific details.

[0077] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0078] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be encompassed therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0079] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0080] Finally, it should be noted that the above description has been provided for purposes of illustration and description. Furthermore, the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to be limiting. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will appreciate that the technical solutions of the present invention may be modified or replaced with equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting the remaining capacity of a lithium-ion battery, characterized in that: include: Collect real-time current data, real-time terminal voltage data and real-time surface temperature data of lithium-ion batteries within a preset time window; Calculating preliminary SOC real-time prediction data based on the real-time current data and the actual battery capacity value of the previous cycle; Calculating the real-time internal resistance data of the battery within the preset time window based on the preliminary SOC real-time prediction data; Inputting the real-time internal resistance data and the real-time surface temperature data of the battery into a battery capacity attenuation evaluation module to obtain a battery capacity attenuation factor; Based on the battery capacity attenuation factor, the battery actual capacity value of the previous cycle is updated to obtain an updated battery actual capacity value, and based on the updated battery actual capacity value, the preliminary SOC real-time prediction data is corrected to obtain battery SOC real-time detection data.

2. The method for detecting the remaining capacity of a lithium-ion battery according to claim 1, wherein: Based on the real-time current data and the actual battery capacity value, preliminary SOC real-time prediction data is calculated, including: The preliminary SOC real-time prediction value at each time point is calculated using the following SOC calculation formula, wherein the SOC calculation formula is as follows: Where Δt represents the sampling interval, I avg represents the average current value within the sampling interval, k represents the current sampling time point, C represents the actual capacity value of the battery, and SOC(·) represents the preliminary SOC real-time prediction value at the corresponding sampling time point.

3. The method for detecting the remaining capacity of a lithium-ion battery according to claim 2, wherein: Calculating the real-time internal resistance data of the battery within the preset time window based on the preliminary SOC real-time prediction data includes: According to SOC-V OCV The curve obtains the battery open circuit voltage data; The real-time internal resistance of the battery at each time point is calculated using the following battery internal resistance calculation formula: Among them, V OCV Indicates the battery open circuit voltage value, V 端 represents the real-time terminal voltage value, I represents the real-time current value, |·| represents the absolute value, R in Indicates the real-time internal resistance of the battery.

4. The method for detecting the remaining capacity of a lithium-ion battery according to claim 1, wherein: Also includes: In response to the real-time terminal voltage value of the lithium-ion battery reaching the charge cut-off voltage or the discharge cut-off voltage, the battery SOC real-time detection data is reset to 100% or 0%.

5. The method for detecting the remaining capacity of a lithium-ion battery according to claim 1, wherein: Inputting the real-time internal resistance data and the real-time surface temperature data of the battery into a battery capacity attenuation evaluation module to obtain a battery capacity attenuation factor includes: respectively extracting the time series fluctuation characteristics of the real-time internal resistance data of the battery and the real-time surface temperature data to obtain a battery internal resistance time series fluctuation characteristic coding vector and a battery surface temperature time series fluctuation characteristic coding vector; Performing time-series interactive response reasoning on the battery internal resistance time-series fluctuation feature coding vector and the battery surface temperature time-series fluctuation feature coding vector to obtain a battery capacity attenuation state feature coding vector; Feature decoding is performed on the battery capacity decay state feature coding vector to obtain the battery capacity decay factor.

6. The method for detecting the remaining capacity of a lithium-ion battery according to claim 5, wherein: Extracting the time series fluctuation characteristics of the real-time internal resistance data of the battery and the real-time surface temperature data respectively to obtain a battery internal resistance time series fluctuation characteristic coding vector and a battery surface temperature time series fluctuation characteristic coding vector, including: The real-time internal resistance data of the battery and the real-time surface temperature data are respectively subjected to time series fluctuation feature extraction based on one-dimensional convolution coding to obtain the time series fluctuation feature coding vector of the battery internal resistance and the time series fluctuation feature coding vector of the battery surface temperature.

7. The method for detecting the remaining capacity of a lithium-ion battery according to claim 6, wherein: Performing time-series interactive response reasoning on the battery internal resistance time-series fluctuation feature coding vector and the battery surface temperature time-series fluctuation feature coding vector to obtain a battery capacity attenuation state feature coding vector, including: Calculating a full-view microscale characteristic response matrix of the battery surface temperature time-series fluctuation characteristic coding vector relative to the battery internal resistance time-series fluctuation characteristic coding vector to obtain a full-view microscale characteristic response matrix of the battery surface temperature-internal resistance time-series fluctuation characteristic; Performing multidimensional feature screening on the full-view microscale feature response matrix of the battery surface temperature-internal resistance time-series fluctuation feature to obtain a sequence distribution of the local response encoding vector of the battery surface temperature-internal resistance time-series fluctuation feature; A global coupling correlation deduction is performed on the sequence distribution of the local response coding vector of the battery surface temperature-internal resistance time series fluctuation characteristic to obtain the battery capacity attenuation state characteristic coding vector.

8. The method for detecting the remaining capacity of a lithium-ion battery according to claim 7, wherein: Multidimensional feature screening is performed on the full-view microscale feature response matrix of the battery surface temperature-internal resistance time series fluctuation feature to obtain a sequence distribution of the local response encoding vector of the battery surface temperature-internal resistance time series fluctuation feature, including: Performing dual-path ground state interactive degenerate optimization on the full-view microscale characteristic response matrix of the battery surface temperature-internal resistance time series fluctuation characteristics to obtain an optimized full-view microscale characteristic response matrix of the battery surface temperature-internal resistance time series fluctuation characteristics; The optimized battery surface temperature-internal resistance timing fluctuation characteristic full-view microscale characteristic response matrix is ​​subjected to multi-dimensional feature screening according to column vectors to obtain the sequence distribution of the battery surface temperature-internal resistance timing fluctuation characteristic local response encoding vector.

9. The method for detecting the remaining capacity of a lithium-ion battery according to claim 8, wherein: Performing global coupling correlation deduction on the sequence distribution of the local response coding vector of the battery surface temperature-internal resistance time series fluctuation characteristic to obtain the battery capacity attenuation state characteristic coding vector, including: The sequence distribution of the local response encoding vector of the battery surface temperature-internal resistance time series fluctuation characteristic is input into the feature response deduction architecture based on the forward LSTM model to obtain the battery capacity attenuation state feature encoding vector.

Images