Lithium Battery SOC-SOH Joint Estimation Method, Device, Equipment and Medium

Abstract

The present invention relates to the technical field of battery energy storage, and in particular to a method, device, equipment and medium for jointly estimating the state of charge (SOC) and state of health (SOH) of a lithium battery, including building a second-order RC equivalent circuit model for the lithium battery, using the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions, and extracting a multi-dimensional battery health factor vector; using a support vector machine regression model to analyze the change trend of the multi-dimensional battery health factor vector to obtain an SOH estimated value; based on the second-order RC equivalent circuit model, using an extended Kalman filter algorithm to perform SOC estimation, and correcting the SOC estimation process through the SOH estimated value to obtain an SOC estimated value. By jointly and highly accurately estimating the SOC-SOH of the lithium battery in the energy storage station, the present invention can effectively reflect the battery aging process and realize high-precision battery state monitoring throughout the operation process and the entire life cycle of the energy storage station.

Description

Technical Field

[0001] The present invention relates to the technical field of battery energy storage, and particularly to a method, device, equipment and medium for jointly estimating the state of charge (SOC) and state of health (SOH) of a lithium battery. Background Art

[0002] The method for jointly estimating the state of charge (SOC) and state of health (SOH) of a lithium battery integrates the internal working principle, performance parameter monitoring, attenuation mechanism analysis and estimation algorithm technology of the lithium battery, aiming to accurately describe the real-time state and long-term health status of the lithium battery. The SOC estimation aims to evaluate the remaining power of the battery in real time. This process is achieved by comprehensively considering multi-dimensional operating parameters such as voltage, current, and temperature, and combining algorithms such as the open-circuit voltage method, Coulomb metering method, and Kalman filter to ensure the accuracy and real-time nature of the power information. The SOH estimation focuses on evaluating the degree of battery aging and its remaining service life. Based on the battery capacity attenuation trend, internal resistance dynamic change and other aging characteristics, it uses model-driven (such as equivalent circuit model) and data-driven (such as machine learning algorithm) methods to comprehensively analyze and evaluate the battery health status. This digital-analog dual-driven evaluation method not only improves the accuracy of the evaluation, but also enhances the adaptability and robustness of the system.

[0003] However, the current methods for jointly estimating SOC and SOH usually adopt technologies such as recursive algorithms, particle filters, and Bayesian estimation, and combine the fusion of real-time monitoring data and historical usage data to build a complex battery behavior model to achieve the synchronous estimation of SOC and SOH. However, when facing battery aging and the resulting changes in model parameters, the current methods for jointly estimating SOC and SOH will show certain limitations. In particular, key parameters such as resistance and capacitance in the equivalent circuit model will continuously fluctuate with the battery aging process and the change of the working environment in practical applications. If the update of these parameters fails to keep up with this dynamic change in a timely manner, it will affect the accuracy and reliability of SOC and SOH estimations.

[0004] In view of the above problems, there is an urgent need to provide a method for jointly estimating SOC and SOH that can adapt to the aging characteristics of the battery, can capture and respond to the changes in the internal state of the battery in real time, provide accurate and reliable SOC and SOH estimation results, and provide solid technical support for the safe and efficient operation of lithium batteries. Summary of the Invention

[0005] The purpose of the present invention is to provide a method, device, equipment and medium for jointly estimating the state of charge (SOC) and state of health (SOH) of a lithium battery. By jointly and highly accurately estimating the SOC and SOH of the lithium batteries in an energy storage station, it can effectively reflect the battery aging process and realize the high-precision state monitoring of the batteries throughout the entire process and full cycle of the operation of the energy storage station.

[0006] To solve the above technical problems, the present invention provides a method, device, equipment and medium for jointly estimating the state of charge (SOC) and state of health (SOH) of a lithium battery.

[0007] In a first aspect, the present invention provides a method for jointly estimating the SOC and SOH of a lithium battery, the method comprising the following steps:

[0008] Model the lithium battery using a second-order RC equivalent circuit model, and simulate the dynamic behavior of the lithium battery under different charge and discharge conditions using the second-order RC equivalent circuit model to extract a multi-dimensional battery health factor vector;

[0009] Analyze the trend of change of the multi-dimensional battery health factor vector using a support vector machine regression model to obtain an SOH estimate value;

[0010] Based on the second-order RC equivalent circuit model, use the extended Kalman filter algorithm to perform SOC estimation, and correct the SOC estimation process through the SOH estimate value to obtain the SOC estimate value.

[0011] In a further implementation, the step of simulating the dynamic behavior of the lithium battery under different charge and discharge conditions using the second-order RC equivalent circuit model to extract a multi-dimensional battery health factor vector includes:

[0012] Simulate the voltage response of the lithium battery under different charge and discharge conditions using the second-order RC equivalent circuit model, divide the charging voltage range during the charging process of the lithium battery into several voltage intervals, and count the interval charging time of each voltage interval;

[0013] Use the sliding window method to obtain the average current within each voltage interval during the battery charging process of the lithium battery, and count the current change rate during the constant voltage stage;

[0014] Extract the battery capacity increment curve from the charge and discharge data of the lithium battery during the charge and discharge process, and extract the peak parameter of the capacity increment curve from the battery capacity increment curve; wherein, the peak parameter of the capacity increment curve includes the maximum peak of the battery capacity increment curve, the time point corresponding to the maximum peak, and the second peak value;

[0015] Use the numerical integration method to calculate the area of the battery capacity increment curve, and calculate the symmetry index of the capacity increment curve according to the area difference between the peak parameter of the capacity increment curve and the areas on the left and right sides of the battery capacity increment curve;

[0016] Construct a multi-dimensional battery health factor vector according to the interval charging time, the average current, the current change rate during the constant voltage stage, and the symmetry index of the capacity increment curve.

[0017] In a further embodiment, the mathematical expression of the symmetry index of the capacity increment curve is as follows:

[0018]

[0019] Wherein,

[0020]

[0021] +

[0022]

[0023]

[0024]

[0025] In the formula, is the symmetry index of the capacity increment curve; A is the total area of the battery capacity increment curve; is the area difference between the areas on the left and right sides of the battery capacity increment curve; is the second peak value of the battery capacity increment curve; is the maximum peak value of the battery capacity increment curve; T is the time length of the charge-discharge process; is the time point corresponding to the maximum peak value; is the area on the left side of the battery capacity increment curve; is the area on the right side of the battery capacity increment curve; is the position index of the maximum peak value; is the battery capacity increment at the time point ; is the battery capacity increment at the next time point ; is the voltage value corresponding to the time point ; is the battery capacity increment at the time point where the maximum peak value appears; is the position index corresponding to the battery capacity increment of the data point; is the position index corresponding to the voltage value of the data point; is the voltage value of the data point corresponding to the previous position index ; n is the total number of data sampling points.

[0026] In a further embodiment, the steps of using the support vector machine regression model to analyze the change trend of the multi-dimensional battery health factor vector and obtain the SOH estimation value include:

[0027] Analyze the changing trend of the multi-dimensional battery health factor vector over time using time series analysis methods, and calculate the change rate of each battery health factor in the multi-dimensional battery health factor vector based on the changing trend of the multi-dimensional battery health factor vector over time to obtain a health factor change rate vector;

[0028] Randomly generate particles to form a particle swarm, where each particle represents a set of support vector machine regression model parameters;

[0029] Randomly initialize the position and velocity of each particle, and use the mean square error of the support vector machine regression model as the fitness function to calculate the fitness of each particle;

[0030] Calculate the adaptive inertia weight according to the health factor change rate vector, and iteratively update the position and velocity of each particle based on the fitness of each particle and the adaptive inertia weight until the preset iterative convergence condition is reached, and output the optimal support vector machine regression model parameters;

[0031] Construct a support vector machine regression model using the optimal support vector machine regression model parameters, and use the support vector machine regression model to predict the health factor change rate vector to obtain the SOH estimation value.

[0032] In a further implementation, the calculation formula for the adaptive inertia weight is:

[0033]

[0034]

[0035] Where,

[0036]

[0037] In the formula, is the adaptive inertia weight at time (t + 1); is the inertia weight at time (t + 1); is the initial value of the inertia weight; is the inertia weight adjustment coefficient; is the change rate of the q-th health factor at time t; h is the number of health factors; is the minimum value of the inertia weight; is the maximum value of the inertia weight; is the battery health status indicator; is the adjustment coefficient of the cumulative effect of the historical health factor change rate; is the time point integration variable; is the current health status adjustment coefficient; is the value of the q-th health factor; is the average value of the q-th health factor; is the standard deviation of the q-th health factor.

[0038] In a further embodiment, the steps of estimating the SOC by using the extended Kalman filter algorithm based on the second-order RC equivalent circuit model and correcting the SOC estimation process through the SOH estimation value to obtain the SOC estimation value include:

[0039] Select the measured battery terminal voltage as the observable according to the second-order RC equivalent circuit model, and establish the observation equation of the extended Kalman filter;

[0040] Input the predefined SOH sample values into the generative adversarial network to generate corresponding augmented SOC-OCV data pairs for each SOH sample value;

[0041] Based on the augmented SOC-OCV data pairs, use the polynomial fitting method to fit the SOC-OCV curves under different SOH estimation value states to obtain the corrected fitting function;

[0042] Correct the original SOC-OCV curve according to the corrected fitting function and the open-circuit voltage of the lithium battery to obtain the voltage correction value;

[0043] Introduce the voltage correction value into the observation equation of the extended Kalman filter to obtain the updated observation equation;

[0044] According to the updated observation equation, use the recursive principle of the extended Kalman filter to estimate the state information of the lithium battery and obtain the SOC estimation value of the lithium battery.

[0045] In a further embodiment, the updated observation equation is:

[0046]

[0047] Wherein,

[0048]

[0049] In the formula, is the observed value at the current moment; is the internal resistance of the battery; is the load current at the k-th moment; is the voltage of the polarization capacitor at the k-th moment; is the observation noise; is the voltage correction value; is the open-circuit voltage of the lithium battery; is the correction function based on the state of health (SOH) of the battery.

[0050] In a second aspect, the present invention provides a lithium battery SOC-SOH combined estimation device, which includes:

[0051] A health factor extraction module, which is used to model the lithium battery with a second-order RC equivalent circuit model, simulate the dynamic behavior of the lithium battery under different charge and discharge conditions by using the second-order RC equivalent circuit model, and extract a multi-dimensional battery health factor vector;

[0052] A state of health estimation module, which is used to analyze the change trend of the multi-dimensional battery health factor vector by using a support vector machine regression model to obtain an SOH estimation value;

[0053] A state of charge estimation module, which is used to perform SOC estimation based on the second-order RC equivalent circuit model by using an extended Kalman filter algorithm, and correct the SOC estimation process through the SOH estimation value to obtain an SOC estimation value.

[0054] In a third aspect, the present invention further provides a computer device, including a processor and a memory. The processor is connected to the memory. The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the computer device executes the steps of implementing the above method.

[0055] In a fourth aspect, the present invention further provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of implementing the above method are realized.

[0056] The present invention provides a lithium battery SOC-SOH combined estimation method, device, equipment and medium. The method models the lithium battery with a second-order RC equivalent circuit model, simulates the dynamic behavior of the lithium battery under different charge and discharge conditions by using the second-order RC equivalent circuit model, and extracts a multi-dimensional battery health factor vector; analyzes the change trend of the multi-dimensional battery health factor vector by using a support vector machine regression model to obtain an SOH estimation value; performs SOC estimation based on the second-order RC equivalent circuit model by using an extended Kalman filter algorithm, and corrects the SOC estimation process through the SOH estimation value to obtain an SOC estimation value. Compared with the prior art, this method uses a support vector machine regression model combined with an adaptive weight adjustment mechanism to perform a combined high-precision estimation of SOC-SOH for the lithium batteries in an energy storage station, can accurately monitor the state of charge of the battery during the charge and discharge process, thereby optimizing the use efficiency and life of the battery, and improving the operation efficiency of the energy storage system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a schematic flow chart of the lithium battery SOC-SOH combined estimation method provided by an embodiment of the present invention;

[0058] Figure 2 It is a schematic diagram of the second-order RC equivalent circuit model provided by an embodiment of the present invention;

[0059] Figure 3 It is a schematic diagram of the generative adversarial network architecture provided by an embodiment of the present invention;

[0060] Figure 4 It is a block diagram of the lithium battery SOC-SOH joint estimation device provided by an embodiment of the present invention;

[0061] Figure 5 It is a schematic diagram of the structure of a computer device provided by an embodiment of the present invention. Detailed implementation manners

[0062] The following specifically clarifies the implementation manners of the present invention in conjunction with the accompanying drawings. The given embodiments are only for illustrative purposes and should not be construed as limiting the present invention. The included drawings are only for reference and explanation and do not constitute a limitation on the protection scope of the present invention's patent, because many changes can be made to the present invention without departing from its spirit and scope.

[0063] Refer to Figure 1 , an embodiment of the present invention provides a method for jointly estimating the SOC and SOH of a lithium battery. As Figure 1 shown, the method includes the following steps:

[0064] S1. Model the lithium battery with a second-order RC equivalent circuit model, and use the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions, and extract a multi-dimensional battery health factor vector.

[0065] In the battery management system (BMS), in this embodiment, a second-order RC equivalent circuit model is constructed to simulate the dynamic behavior of the lithium battery in order to estimate the SOC (remaining charge) of the battery, predict the battery life, perform fault diagnosis and other data. The second-order RC equivalent circuit model includes an ideal voltage source , an ohmic resistance and two RC parallel circuits. In Figure 2 , the ideal voltage source represents the open-circuit voltage of the battery, that is, the voltage of the battery when there is no load; the ohmic resistance represents the internal resistance of the battery (the ohmic resistance inside the battery), which directly affects the voltage drop of the battery during the charge and discharge process; the two RC parallel circuits include the battery electrochemical polarization internal resistance and capacitance , and the concentration polarization effect internal resistance and capacitance . Among them, the battery electrochemical polarization internal resistance and capacitance Reflects the voltage change inside the battery due to electrochemical reactions, concentration polarization effect internal resistance and capacitance Reflects the voltage change inside the battery due to ion concentration changes; Represents the battery terminal voltage; Represents the input current.

[0066] This embodiment describes the voltage and current relationship of the battery in the time domain according to circuit principles. The electrical characteristic equation of the second-order RC equivalent circuit model in the time domain is specifically:

[0067]

[0068] In the formula, is the battery terminal voltage; is the open-circuit voltage is a function of SOC; is the parallel circuit voltage across both ends; is the parallel circuit voltage across both ends; is the input current at time t; is the battery internal resistance.

[0069] Converting the electrical characteristic equation of the second-order RC equivalent circuit model in the time domain into a transfer function in the frequency domain facilitates frequency-domain analysis and parameter identification. Specifically:

[0070]

[0071]

[0072] After arrangement, it can be obtained:

[0073]

[0074] In the formula, is the difference between the terminal voltage and the open-circuit voltage in the complex frequency domain; is the open-circuit voltage in the complex frequency domain; is the battery terminal voltage in the complex frequency domain; s is the variable in the complex frequency domain; is the Laplace transform of the load current; is the transfer function in the frequency domain, representing the relationship between the difference between the terminal voltage and the open-circuit voltage and the battery terminal current.

[0075] In order to apply in an actual digital system, this embodiment needs to convert the continuous-time transfer function into a discrete-time transfer function. Specifically, using the bilinear transformation formula , converting the continuous domain to the discrete domain, where T is the sampling time; z represents the discrete domain variable; this embodiment sets 、 , a discrete-time transfer function is obtained :

[0076]

[0077] In the formula, k is a variable; is the coefficient obtained through bilinear transformation, which is related to the model parameters and sampling time, as shown in the following formula:

[0078]

[0079] Define parameters a, b, c, d, e, as shown in the following formula:

[0080]

[0081] After simplification, it can be obtained:

[0082]

[0083]

[0084] In this embodiment, by constructing a second-order RC equivalent circuit model and using circuit principles to derive the electrical characteristic equations in the time domain and frequency domain, and finally converting them into a discrete-time transfer function through bilinear transformation. This second-order RC equivalent circuit model not only reflects the internal resistance of the battery, but also takes into account the effects of electrochemical polarization and concentration polarization, so as to be able to more accurately simulate the dynamic behavior of lithium batteries under different charge and discharge conditions, including voltage response, current change, etc., providing a reliable data basis for subsequent extraction of health factors and SOH estimation. In this embodiment, the multi-dimensional battery health factor vector includes interval charge time, average current, current change rate in the constant voltage stage, and symmetry index of the capacity increment curve. The steps of using the second-order RC equivalent circuit model to simulate the dynamic behavior of lithium batteries under different charge and discharge conditions and extract the multi-dimensional battery health factor vector include:

[0085] Use the second-order RC equivalent circuit model to simulate the voltage response of the lithium battery under different charge and discharge conditions, divide the charging voltage range during the charging process of the lithium battery into several voltage intervals, and count the interval charge time of each voltage interval;

[0086] Use the sliding window method to obtain the average current in each voltage interval during the battery charging process of the lithium battery, and count the current change rate in the constant voltage stage;

[0087] Extract the battery capacity increment curve from the charge and discharge data of the lithium battery during the charge and discharge process, and extract the peak parameters of the capacity increment curve from the battery capacity increment curve; wherein, the peak parameters of the capacity increment curve include the maximum peak of the battery capacity increment curve, the time point corresponding to the maximum peak, and the second peak value.

[0088] Use the numerical integration method to calculate the area of the battery capacity increment curve, and calculate the symmetry index of the capacity increment curve according to the area difference between the peak parameters of the capacity increment curve and the areas on the left and right sides of the battery capacity increment curve.

[0089] Construct a multi-dimensional battery health factor vector according to the interval charging time, the average current, the current change rate in the constant voltage stage, and the symmetry index of the capacity increment curve.

[0090] Specifically, since the discharge working condition of the battery cannot be predicted during the actual working process, in contrast, its charging working condition is relatively stable, fixed and predictable, and it can be considered that the health state (SOH) of the battery remains basically unchanged within a single charge and discharge cycle. Therefore, in this embodiment, the characteristic representations related to battery aging can be extracted from the charging working condition data through experimental results. As the charge and discharge cycles proceed, within the same charging time period, the fluctuation range of the battery voltage gradually increases. This change reveals that the time required for the battery to reach the same voltage difference during charging is gradually shortening, that is, the response speed of the battery to charging is accelerating. There is a positive correlation between the charging time and the battery capacity. The charging time can be used as an important reference basis for evaluating the reduction of battery capacity and the degree of battery aging. Therefore, in this embodiment, the charging time from the starting voltage to the constant current cut-off voltage is selected as the health factor, and at the same time, the charging time is refined and the charging time in different voltage intervals is statistically segmented as an independent health factor. The specific implementation process is as follows:

[0091] In this embodiment, the charging voltage range of the lithium battery during charging is divided into several voltage intervals. For example, assume that the charging voltage range of the battery during charging is to , this embodiment can divide it into M voltage intervals, and the voltage range of each voltage interval is , where is the starting voltage of each voltage interval, is the end voltage of each voltage interval. Meanwhile, in this embodiment, the interval charging time of each voltage interval is counted. The voltage interval and the interval charging time of each voltage interval are used to describe the change of the charging performance of the battery at different voltages. This embodiment can set a reasonable number of intervals. While considering the computational complexity and data storage volume in practical applications, ensure that the voltage range of each interval is small enough to capture the subtle changes in battery performance. For example, the voltage range is divided into five intervals, and the interval charging times of these five voltage intervals are all used as health factors.

[0092] Since the charging cut-off current is the same, an increase in charging time means that the charging current decline rate of the battery in the later stage of aging slows down during the constant voltage stage. There is a negative correlation between the average current and the battery capacity within the same time. Considering the possible incompleteness of the charging cycle in practical engineering applications, this embodiment uses the sliding window method to obtain the average current of the lithium battery within each voltage interval during the battery charging process as a health factor. In addition, to improve the extraction of the current health factor, in addition to using the average value of the current, this embodiment proposes to use the change rate of the current (i.e., the slope of the current decline) during the constant voltage stage as a health factor, so that the support vector machine regression model can more sensitively capture the change of the charging behavior of the battery during the constant voltage stage. The process of obtaining the current change rate is specifically as follows:

[0093] The charging process of the lithium battery is usually divided into three stages: constant current charging (CC), constant voltage charging (CV), and floating charging stage. During the constant current stage, the battery is charged with the maximum current until the voltage reaches a preset voltage threshold, and then enters the constant voltage stage. At this time, the voltage remains unchanged and the current gradually decreases until the current drops to a certain low value and enters the floating charging stage. In this embodiment, during the battery charging process, when entering the constant voltage stage, current data is periodically collected at a preset sampling time interval, and then linear regression is used to fit the collected current data points to reveal the linear pattern of the current change with time, so as to accurately calculate the change rate (i.e., the slope) of the current within the constant voltage stage. For example, if the linear regression model obtained by fitting is expressed as I(t)= t + , where represents the change rate (slope) of the current with time; t represents time; I(t) is the predicted current value at a given time t; is the intercept term; through the above processing, this embodiment can quantify the change speed of the battery charging current within the constant voltage stage, which is of great significance for predicting the remaining charging time and other aspects.

[0094] The battery capacity increment curve has unique shape, height, and position characteristics, which reflect the electrochemical reactions in the battery during charge and discharge cycles. In particular, there are two distinct peaks on this curve, and these two peaks directly reflect the rate at which the battery capacity changes sharply with a small change in voltage under the current voltage conditions, that is, the significant change points of the slope. As the battery aging process intensifies, the heights of the two peaks gradually decrease, indicating a decline in the battery's ability to store and release energy. At the same time, the voltage positions corresponding to the peaks gradually shift towards higher voltage regions, further revealing the changes in the internal electrochemical characteristics of the battery with aging. Therefore, in view of the close correlation or "following nature" between the two peaks, in order to avoid introducing the problem of multicollinearity in subsequent data analysis, which may affect the effectiveness and accuracy of the model, in this embodiment, it is considered to extract the peak parameters of the capacity increment curve from the battery capacity increment curve. Among them, the peak parameters of the capacity increment curve include the maximum peak of the battery capacity increment curve, the time point corresponding to the maximum peak, and the second peak value, so as to effectively characterize the aging state of the battery. In addition, to further improve the extraction method of the battery health factor, in addition to extracting the peak parameters of the capacity increment curve of the capacity increment curve, this embodiment proposes to extract the symmetry index of the capacity increment curve. These characteristics can more comprehensively reflect the capacity change and aging state of the battery. The following is the specific implementation process:

[0095] In this embodiment, a numerical integration method is used to calculate the area of the battery capacity increment curve. Assuming the curve data points are , the calculation formula for the total area of the battery capacity increment curve is:

[0096]

[0097] In the formula, A is the area of the battery capacity increment curve; is the battery capacity increment at time point , that is, the amount of electricity increased by the battery at time point relative to the previous time point ; is the battery capacity increment at time point ; is the voltage value corresponding to time point ; is the voltage value corresponding to time point ; n is the total number of data sampling points.

[0098] Then, in this embodiment, the symmetry is measured according to the area difference between the peak parameters of the capacity increment curve and the areas on the left and right sides of the battery capacity increment curve. Among them, the calculation formulas for the areas on the left and right sides of the battery capacity increment curve are:

[0099] +

[0100]

[0101]

[0102]

[0103] The specific calculation formula for the symmetry index of the capacity increment curve is as follows:

[0104]

[0105]

[0106] In the formula, is the symmetry index of the capacity increment curve; A is the total area of the battery capacity increment curve; is the area difference between the areas on the left and right sides of the battery capacity increment curve; is the second peak value of the battery capacity increment curve. In this embodiment, the area difference and the second peak are used as penalty factors for the symmetry of the capacity increment curve; is the maximum peak value of the battery capacity increment curve, The smaller the ratio, the more uniform the peak distribution of the capacity increment curve and the better the symmetry of the capacity increment curve; T is the time length of the charge-discharge process; is the time point corresponding to the maximum peak value, can be used to measure the centrality of the peak position. The more centered the peak position, the better the symmetry of the capacity increment curve; is the area on the left side of the battery capacity increment curve; is the area on the right side of the battery capacity increment curve; is the position index of the maximum peak value, that is, the data point index corresponding to the highest point on the battery capacity increment curve; is the battery capacity increment at the time point ; is the next time point ; is the time point ; is the battery capacity increment at the time point where the maximum peak value appears; is the position index ; is the battery capacity increment at the time point is the position index ; is the voltage value of the data point corresponding to the previous position index ; n is the total number of data sampling points; To calculate the area from to where, if is not a sampling point, the incremental battery capacity at can be estimated by interpolation; if is a sampling point, the incremental battery capacity of this sampling point (data point) is directly used, and the calculation of the right - hand area is similar; To calculate the area from to .

[0107] S2. Use a multi - dimensional support vector machine regression model to analyze the changing trend of the multi - dimensional battery health factor vector to obtain the SOH estimated value.

[0108] In this embodiment, the steps of using the support vector machine regression model to analyze the changing trend of the multi - dimensional battery health factor vector to obtain the SOH estimated value include:

[0109] Use the time - series analysis method to analyze the changing trend of the multi - dimensional battery health factor vector over time, and calculate the change rate of each battery health factor in the multi - dimensional battery health factor vector according to the changing trend of the multi - dimensional battery health factor vector over time to obtain the health factor change rate vector; the health factor change rate vector includes the change rate of the interval charging time, the change rate of the average current, the change rate of the current in the constant - voltage stage, and the change rate of the symmetry index of the capacity increment curve;

[0110] Randomly generate particles to form a particle swarm, and each particle represents a set of support vector machine regression model parameters;

[0111] Randomly initialize the position and velocity for each particle, and use the mean square error of the support vector machine regression model as the fitness function to calculate the fitness of each particle;

[0112] Calculate the adaptive inertia weight according to the health factor change rate vector, and iteratively update the position and velocity of each particle according to the fitness of each particle and the adaptive inertia weight until the preset iteration convergence condition is reached, and output the optimal support vector machine regression model parameters;

[0113] Use the optimal support vector machine regression model parameters to construct a support vector machine regression model, and use the support vector machine regression model to predict the health factor change rate vector to obtain the SOH estimated value.

[0114] Specifically, to optimize the performance of the SVR model (Support Vector Regression model), in the process of estimating the state of health (SOH) of the battery in this embodiment, by combining the Particle Swarm Optimization (PSO) algorithm and an adaptive weight adjustment mechanism, it can more flexibly respond to the dynamic changes of health factors, thereby improving the accuracy of SOH estimation. The specific implementation process is as follows:

[0115] In this embodiment, the data of each health factor at time point t is collected, and the change rate of each health factor is calculated using time series analysis methods. These change rates reflect the trend of the battery state of health evolving over time and are an important basis for subsequent adaptive weight adjustment;

[0116] The PSO algorithm searches for the optimal solution in the solution space by simulating swarm intelligence behavior. The prediction error (such as the mean square error MSE) of the SVR model is used as the fitness function. Each particle updates its position and velocity in the search space based on individual experience and swarm experience to approach the optimal solution. Through iterative search, the parameter combination that minimizes the prediction error of the SVR model is finally found, that is, the optimal SVR model parameters. The update formulas for the velocity and position of the particle are:

[0117]

[0118] In the formula, is the velocity of particle at moment, is the position of particle at moment, is the individual optimal position of particle , is the global optimal position, is the inertia weight, and are learning factors, and are random numbers between 0 and 1.

[0119] To improve the response sensitivity of the PSO algorithm to the changes in the battery state of health, this embodiment introduces an adaptive weight mechanism, which dynamically adjusts the adaptive inertia weight according to the change rate of the health factor. The change rate vector of the health factor is defined as , where each component of the health factor change rate vector represents the change rate of the corresponding health factor. The initial value of the inertia weight is set to , and the adjustment range is set to . The calculation formula for the adaptive inertia weight is:

[0120]

[0121] Among them,

[0122]

[0123] Meanwhile, ensure that the inertia adaptive weight is within the effective range:

[0124]

[0125] In the formula, is the adaptive inertia weight after adjustment within the effective range at the (t + 1)th moment; is the inertia weight at the (t + 1)th moment; is the initial value of the inertia weight; is the inertia weight adjustment coefficient; is the change rate of the qth health factor at the tth moment; h is the number of health factors; is the minimum value of the inertia weight; is the maximum value of the inertia weight. It should be noted that in this embodiment, ensures that the adaptive inertia weight is not lower than the minimum value , and through ensures that the adaptive inertia weight is not higher than the maximum value ; is the battery health state index; is the adjustment coefficient of the cumulative effect of the historical health factor change rate, which is used to adjust the influence of the cumulative historical health factor change rate on the current inertia weight; is the time point integration variable; represents the integral of the absolute values of the change rates of all health factors from the 0th moment to the tth moment, which reflects the cumulative effect of the health factor change rate to capture the long-term trend and better analyze the change trend of the health state, rather than being limited to considering short-term changes; is the current health state adjustment coefficient; is the value of the qth health factor; is the average value of the qth health factor; is the standard deviation of the qth health factor.

[0126] The adaptive weight adjustment mechanism adopted in this embodiment comprehensively considers the current and historical changes of the health factor, enabling the algorithm to more accurately predict the SOH of the battery when facing different battery health state changes, thereby improving the prediction accuracy of the model. At the same time, by dynamically adjusting the inertia weight, the movement of the particles in the search space becomes more flexible, which helps to avoid falling into local optima and improves the efficiency of global search. In summary, through the above steps, this embodiment constructs an SVR model based on multi-dimensional health factors to predict SOH. At the same time, by introducing the adaptive weight adjustment mechanism of the PSO algorithm, the SVR model can more accurately estimate the SOH of the battery, thus providing more reliable data support for the battery management system.

[0127] S3. Based on the second-order RC equivalent circuit model, use the extended Kalman filter algorithm to estimate the SOC, and correct the SOC estimation process through the SOH estimation value to obtain the SOC estimation value.

[0128] In this embodiment, the steps of estimating the SOC based on the second-order RC equivalent circuit model using the extended Kalman filter algorithm and correcting the SOC estimation process through the SOH estimation value to obtain the SOC estimation value include:

[0129] According to the second-order RC equivalent circuit model, select the SOC of the battery and the voltage of the polarization capacitor as state variables using the ampere-hour method principle of battery SOC, and establish a system state equation;

[0130] Select the measured battery terminal voltage as the observable quantity and establish an observation equation; the observation equation is used to describe the relationship between the battery terminal voltage and the state variables;

[0131] Discretize the continuous second-order RC equivalent circuit model to obtain a discrete nonlinear equation, and in each iteration, linearize the discrete nonlinear equation to obtain the linearized forms of the system state equation and the observation equation;

[0132] According to the linearized forms of the system state equation and the observation equation, determine the extended Kalman filter matrix; the extended Kalman filter matrix includes the state transition matrix and the observation matrix, etc.;

[0133] According to the Kalman filter recursion principle and the extended Kalman filter matrix, obtain the extended Kalman filter recursion formula based on the second-order RC equivalent circuit model;

[0134] Use the polynomial fitting method to fit the SOC-OCV curve under different SOH estimation value states to obtain a corrected fitting function;

[0135] According to the corrected fitting function and the open-circuit voltage of the lithium battery, correct the original SOC-OCV curve to obtain a voltage correction value;

[0136] Introduce the voltage correction value into the observation equation of the extended Kalman filter to obtain the updated observation equation;

[0137] According to the updated observation equation and the extended Kalman filter recurrence formula, use the recurrence principle of the extended Kalman filter to estimate the state information of the lithium battery and obtain the SOC estimation value of the lithium battery.

[0138] Specifically, the battery SOC estimation model is based on the equivalent circuit model, the second-order RC battery model, which includes components such as polarization capacitance and polarization resistance, and can better simulate the dynamic characteristics of the battery during charging and discharging. In this embodiment, the SOC (percentage of remaining battery charge) of the battery and the voltage of the polarization capacitance are selected as the state variables of the system, and the system state equation is established through the second-order RC equivalent circuit model and the ampere-hour method principle. The system state equation relates the state variables to parameters such as the current of the battery, the polarization resistance, and the Coulomb efficiency. The specific system state equation is:

[0139]

[0140]

[0141] In the formula, is the state vector at time k; is the state transition matrix at time k; is the sampling time; is the voltage of the polarization capacitance at time k; is the polarization capacitance; is the polarization resistance; is the rated capacity of the battery; is the Coulomb efficiency; is the current at time (k - 1); is the load current at time k; and is the system noise; is the observation noise.

[0142] Then, in this embodiment, the battery terminal voltage is selected as the observable quantity to establish an observation equation, which describes the relationship between the battery terminal voltage and the state variables. Since the state equation and the observation equation may be non-linear, in each iteration of this embodiment, the Jacobian matrix is calculated to linearize these equations and obtain the linearized forms of the state equation and the observation equation. Some linear transformations are also required during the linearization process of the non-linear discrete system. The matrices required for the linear transformation process of the Kalman filter:

[0143]

[0144] In the formula, is the partial derivative of the battery terminal voltage with respect to SOC; is the control input matrix; is the observation matrix.

[0145] According to the results of linearization, determine the state transition matrix, control input matrix, observation matrix, etc. required by the Kalman filter. According to the recursive principle of the Kalman filter, combined with the above equations and matrices, the recursive formula of the extended Kalman filter based on the second-order RC model is derived:

[0146]

[0147] In the formula, is the initial state estimate; is the initial error covariance matrix; is the predicted state estimate of the state at time k based on the state at time k - 1; is the control input at time k; is the predicted error covariance matrix; is the process noise distribution matrix; is the process noise covariance matrix; is the Kalman gain matrix; is the measurement noise covariance; is the updated state estimate at time k; is the new measurement value, is the measurement value predicted from the state estimate, is the updated posterior estimate error covariance matrix.

[0148] It can be seen from the above steps that in this embodiment, the predicted state estimate and the predicted error covariance matrix are calculated using the state equation and the state estimate at the previous moment. Using the new measurement value , combined with the Kalman gain , then the state estimate and the error covariance matrix are updated. Repeat the above prediction and update steps until a satisfactory SOC estimation result is obtained.

[0149] In the process of estimating the State of Charge (SOC), considering that the difference in the SOC-OCV (Open Circuit Voltage) curve under different States of Health (SOH) has a direct impact on the accuracy of parameter update and affects the accuracy of SOC estimation. Therefore, in this embodiment, the SOC-OCV relationship is corrected by the real-time updated SOH value, which can more accurately optimize the SOC estimation result. In this embodiment, the SOC-OCV curve is dynamically adjusted according to the SOH value to reflect the true voltage characteristics under different health states. The specific implementation process is as follows:

[0150] First, this embodiment selects batteries in different SOH states for experiments. For example, 80%, 90%, and 100% are selected as representative values of the health factor. Then, under each selected SOH state, by controlling the charge and discharge process of the battery, the OCV values corresponding to different SOCs are measured and recorded to form SOC-OCV data pairs. Next, this embodiment uses the polynomial fitting method to fit the SOC-OCV data under each SOH state to obtain the SOC-OCV curves under different SOH states. The fitting results under different SOH states are comprehensively analyzed to identify the influence law of SOH on the SOC-OCV relationship, thereby constructing a correction function f(SOH) that can reflect the influence of SOH on SOC-OCV. The open circuit voltage is dynamically adjusted according to the correction function f(SOH) to improve the accuracy of SOC estimation. The corrected open circuit voltage is the SOC-OCV curve obtained by dynamically adjusting the original open circuit voltage after considering the battery health state (SOH). The corrected open circuit voltage more accurately reflects the actual voltage of the battery in the current health state. The mathematical expression of the voltage correction value is:

[0151]

[0152] In the formula, is the voltage correction value; is the original open circuit voltage of the battery; is the correction function based on the battery health state (SOH).

[0153] Although the correction function can be initially established through the above steps, relying solely on the experimental data in a limited number of SOH states may lead to limited accuracy of the correction function and fail to ensure high precision. Therefore, in the estimation of the state of health (SOH) of the battery, this embodiment uses a GAN (Generative Adversarial Network) to augment the SOC-OCV (State of Charge-Open Circuit Voltage) dataset to address the problem of insufficient experimental data. The GAN can learn the distribution characteristics of the existing data and generate high-quality new data samples, thereby enriching the SOC-OCV dataset without increasing the actual experimental burden. With the dataset augmented by the GAN, the correction function f(SOH) can be further trained and optimized to more accurately reflect the impact of SOH on the SOC-OCV curve, ultimately improving the accuracy of SOC estimation. Figure 3 As shown in the schematic diagram of the generative adversarial network architecture, when using the generative adversarial network (GAN) to generate SOC-OCV data, the generator receives a random noise vector as input and generates SOC-OCV data. In this embodiment, the generator consists of multiple fully connected layers, and maps the input random noise vector to the SOC-OCV space through a multi-layer neural network to generate realistic data samples. The ReLU activation function is used between layers to enhance the non-linear processing ability; the discriminator model receives the SOC-OCV data and the corresponding SOH value as input and judges the authenticity (true or false) of the data. In this embodiment, the discriminator model consists of fully connected layers, maps the input SOC, OCV, and SOH to the probability value of authenticity (0 or 1, representing false or true), and the sigmoid activation function is used in the last layer to output the authenticity probability value between 0 and 1. In this embodiment, the number of training epochs and the number of iterations in each epoch are set. The specific GAN training process is as follows:

[0154] Train the discriminator: Randomly extract a batch of SOC-OCV-SOH data pairs from the real data as real samples; generate an equal number of fake samples as the real samples. Input the random noise and SOH into the generator to generate SOC-OCV data pairs; input the real samples and fake samples into the discriminator to classify the real samples and fake samples through the discriminator (the label of the real sample is 1, and the label of the fake sample is 0); calculate the classification loss of the discriminator and update the parameters of the discriminator so that it can better distinguish between real samples and fake samples;

[0155] Aiming to generate fake samples that the discriminator cannot distinguish, train the generator: Generate fake samples from the random noise, set the label of this batch of fake samples to 1 (regarded as real samples), and input them into the discriminator for classification; calculate the loss of the generator (the difference between the probability output by the discriminator and the real label 1), and update the parameters of the generator so that it generates more realistic SOC-OCV data;

[0156] Alternately train the generator and the discriminator: During each training process, alternately train the discriminator and the generator.

[0157] In this embodiment, a range of SOH samples can be selected, for example, from 0.8 to 1.0, and multiple SOH values (e.g., 10 points) are evenly selected within this range. For each SOH sample, a corresponding SOC-OCV data pair is generated by inputting random noise into the generator. The generated data is de-normalized and restored to the original data range. Using the newly generated SOC-OCV data and combining with the original SOC-OCV data, the correction function f(SOH) is established. Through the above steps, this embodiment can effectively solve the problem of accuracy decline caused by SOH changes during the SOC estimation process through the dynamic adjustment and dataset expansion strategy of the SOC-OCV curve based on the SOH value, improve the accuracy of battery health state estimation, and provide rich dataset support for subsequent battery state estimation or performance analysis.

[0158] This embodiment introduces a correction term based on the SOH value into the extended Kalman filter algorithm , and in the recurrence formula of the Kalman filter, the corrected SOC-OCV curve is used for estimation:

[0159]

[0160] In the formula, is the initial state estimate of the state at time k based on the state at time k−1.

[0161] The embodiment of the present invention provides a method for jointly estimating the SOC and SOH of a lithium battery. The method models the lithium battery using a second-order RC equivalent circuit model, and uses the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions to extract a multi-dimensional battery health factor vector; uses a support vector machine regression model to analyze the change trend of the multi-dimensional battery health factor vector to obtain an SOH estimate value; based on the second-order RC equivalent circuit model, uses the extended Kalman filter algorithm to estimate the SOC, and corrects the SOC estimation process through the SOH estimate value to obtain the SOC estimate value. Compared with the prior art, this method uses a support vector machine regression model combined with an adaptive weight adjustment mechanism, and at the same time uses the extended Kalman filter to jointly and highly accurately estimate the SOC and SOH of the lithium battery in the energy storage station, and can accurately monitor the state of charge of the battery during the charge and discharge process, thereby optimizing the battery usage efficiency and lifespan and improving the operation efficiency of the energy storage system.

[0162] It should be noted that the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0163] In one embodiment, as Figure 4 shown, an embodiment of the present invention provides a lithium battery SOC-SOH joint estimation device, and the device includes:

[0164] A health factor extraction module 101, configured to perform a second-order RC equivalent circuit model modeling on a lithium battery, and use the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions, and extract a multi-dimensional battery health factor vector;

[0165] A state of health estimation module 102, configured to perform a trend analysis on the multi-dimensional battery health factor vector by using a support vector machine regression model to obtain an SOH estimation value;

[0166] A state of charge estimation module 103, configured to perform SOC estimation based on the second-order RC equivalent circuit model by using an extended Kalman filter algorithm, and correct the SOC estimation process through the SOH estimation value to obtain an SOC estimation value.

[0167] For the specific limitations on a lithium battery SOC-SOH joint estimation device, reference can be made to the above limitations on a lithium battery SOC-SOH joint estimation method, which will not be elaborated here. Those of ordinary skill in the art can realize that, in combination with the various modules and steps described in the embodiments disclosed in the present application, they can be implemented in hardware, software, or a combination of both. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0168] An embodiment of the present invention provides a lithium battery SOC-SOH joint estimation device. The device performs a second-order RC equivalent circuit model modeling on a lithium battery through a health factor extraction module, and uses the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions, and extracts a multi-dimensional battery health factor vector; performs a trend analysis on the multi-dimensional battery health factor vector through a state of health estimation module to obtain an SOH estimation value; and the state of charge estimation module uses an extended Kalman filter algorithm to perform SOC estimation, and corrects the SOC estimation process through the SOH estimation value to obtain an SOC estimation value. Compared with the prior art, the device uses a support vector machine regression model combined with an adaptive weight adjustment mechanism, and at the same time uses an extended Kalman filter to perform a joint high-precision estimation of SOC-SOH on the lithium battery of an energy storage station, and can accurately monitor the state of charge of the battery during the charge and discharge process, thereby optimizing the use efficiency and life of the battery and improving the operation efficiency of the energy storage system.

[0169] Figure 5A computer device provided by an embodiment of the present invention includes a memory, a processor, and a transceiver, which are connected through a bus; the memory is used to store a set of computer program instructions and data, and can transmit the stored data to the processor, and the processor can execute the program instructions stored in the memory to execute the steps of the above method.

[0170] Among them, the memory may include a volatile memory or a non-volatile memory, or may include both a volatile and a non-volatile memory; the processor may be a central processing unit, a microprocessor, an application-specific integrated circuit, a programmable logic device, or a combination thereof. By way of example but not limitation, the above programmable logic device may be a complex programmable logic device, a field programmable gate array, a generic array logic, or any combination thereof.

[0171] In addition, the memory may be a physically independent unit or may be integrated with the processor.

[0172] Those of ordinary skill in the art can understand that Figure 5 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have the same component arrangement.

[0173] In one embodiment, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0174] A lithium battery SOC-SOH joint estimation method, device, equipment and medium provided by an embodiment of the present invention. A lithium battery SOC-SOH joint estimation method uses a support vector machine regression (SVR) model combined with an adaptive weight adjustment mechanism to predict the change rate of the battery health factor, so as to estimate the SOH of the battery. At the same time, high-precision estimation of the state of charge of the battery during charging and discharging is achieved through an extended Kalman filter, which can not only improve the operation efficiency and safety of the batteries in the energy storage station, but also reduce the maintenance cost and extend the service life of the battery, realizing high-precision state monitoring of the batteries throughout the whole process and the whole cycle of the operation of the energy storage station.

[0175] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as an SSD), etc.

[0176] Those skilled in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0177] The above embodiments only represent several preferred implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and substitutions can be made, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for jointly estimating the state of charge (SOC) and state of health (SOH) of a lithium battery, characterized in that, It includes the following steps: Build a second-order RC equivalent circuit model for the lithium battery, and use the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions, and extract a multi-dimensional battery health factor vector; Use a support vector machine regression model to analyze the change trend of the multi-dimensional battery health factor vector to obtain an SOH estimation value; Based on the second-order RC equivalent circuit model, use the extended Kalman filter algorithm to perform SOC estimation, and correct the SOC estimation process through the SOH estimation value to obtain the SOC estimation value; Among them, the step of using the second-order RC equivalent circuit model to simulate the dynamic behavior of the lithium battery under different charge and discharge conditions and extract the multi-dimensional battery health factor vector includes: Use the second-order RC equivalent circuit model to simulate the voltage response of the lithium battery under different charge and discharge conditions, divide the charging voltage range during the charging process of the lithium battery into several voltage intervals, and count the interval charging time of each voltage interval; Use the sliding window method to obtain the average current in each voltage interval during the battery charging process of the lithium battery, and count the current change rate during the constant voltage stage; Extract the battery capacity increment curve according to the charge and discharge data of the lithium battery during the charge and discharge process, and extract the peak parameter of the capacity increment curve from the battery capacity increment curve; among them, the peak parameter of the capacity increment curve includes the maximum peak of the battery capacity increment curve, the time point corresponding to the maximum peak, and the second peak value; Use the numerical integration method to calculate the area of the battery capacity increment curve, and calculate the symmetry index of the capacity increment curve according to the area difference between the peak parameter of the capacity increment curve and the areas on the left and right sides of the battery capacity increment curve; Construct a multi-dimensional battery health factor vector according to the interval charging time, the average current, the current change rate during the constant voltage stage, and the symmetry index of the capacity increment curve.

2. The method for jointly estimating the SOC and SOH of a lithium battery according to claim 1, wherein, The mathematical expression of the symmetry index of the capacity increment curve is: Where + Wherein, is the symmetry index of the capacity increment curve; A is the total area of the battery capacity increment curve; is the area difference between the areas on the left and right sides of the battery capacity increment curve; is the second peak value of the battery capacity increment curve; is the maximum peak value of the battery capacity increment curve; T is the time length of the charge-discharge process; is the time point corresponding to the maximum peak value; is the area on the left side of the battery capacity increment curve; is the area on the right side of the battery capacity increment curve; is the position index of the maximum peak value; is at the time point the battery capacity increment; is the next time point the battery capacity increment; is the time point the corresponding voltage value; is the time point when the maximum peak appears the battery capacity increment; is the position index the battery capacity increment of the corresponding data point; is the position index the voltage value of the corresponding data point; is the previous position index the voltage value of the corresponding data point; n is the total number of data sampling points.

3. The SOC-SOH joint estimation method for a lithium battery according to claim 1, wherein The step of using the support vector machine regression model to analyze the change trend of the multi-dimensional battery health factor vector to obtain the SOH estimation value includes: Use the time series analysis method to analyze the change trend of the multi-dimensional battery health factor vector over time, and calculate the change rate of each battery health factor in the multi-dimensional battery health factor vector according to the change trend of the multi-dimensional battery health factor vector over time to obtain the health factor change rate vector; Randomly generate particles to form a particle swarm, and each particle represents a set of support vector machine regression model parameters; Randomly initialize the position and velocity of each particle, and use the mean square error of the support vector machine regression model as the fitness function to calculate the fitness of each particle; Calculate the adaptive inertia weight according to the health factor change rate vector, and iteratively update the position and velocity of each particle according to the fitness of each particle and the adaptive inertia weight until the preset iteration convergence condition is reached, and output the optimal support vector machine regression model parameters; Use the optimal support vector machine regression model parameters to construct a support vector machine regression model, and use the support vector machine regression model to predict the health factor change rate vector to obtain the SOH estimation value.

4. The method for jointly estimating the SOC and SOH of a lithium battery according to claim 3, wherein, The calculation formula of the adaptive inertia weight is as follows: Wherein, Wherein, is the adaptive inertia weight at the (t + 1)th moment; is the inertia weight at the (t + 1)th moment; is the initial value of the inertia weight; is the inertia weight adjustment coefficient; is the change rate of the qth health factor at the tth moment; h is the number of health factors; is the minimum value of the inertia weight; is the maximum value of the inertia weight; is the battery health state indicator; is the adjustment coefficient of the cumulative effect of the historical health factor change rate; is the time point integration variable; is the current health state adjustment coefficient; is the value of the qth health factor; is the average value of the qth health factor; is the standard deviation of the qth health factor.

5. The method for jointly estimating the SOC and SOH of a lithium battery according to claim 1, characterized in that, The steps of estimating the SOC based on the second-order RC equivalent circuit model by using the extended Kalman filter algorithm and correcting the SOC estimation process through the SOH estimation value to obtain the SOC estimation value include: Select the measured battery terminal voltage as the observable according to the second-order RC equivalent circuit model, and establish the observation equation of the extended Kalman filter; Input the predefined SOH sample values into the generative adversarial network to generate corresponding augmented SOC-OCV data pairs for each SOH sample value; Based on the augmented SOC-OCV data pairs, use the polynomial fitting method to fit the SOC-OCV curves under different SOH estimation value states to obtain the correction fitting function; Correct the original SOC-OCV curve according to the correction fitting function and the open-circuit voltage of the lithium battery to obtain the voltage correction value; Introduce the voltage correction value into the observation equation of the extended Kalman filter to obtain the updated observation equation; According to the updated observation equation, use the recursive principle of the extended Kalman filter to estimate the state information of the lithium battery and obtain the SOC estimation value of the lithium battery.

6. The method for jointly estimating the SOC and SOH of a lithium battery according to claim 5, wherein The updated observation equation is: Wherein, Wherein, is the observed value at the current moment; is the internal resistance of the battery; is the load current at the k-th moment; is the voltage of the polarization capacitor at the k-th moment; is the observation noise; is the voltage correction value; is the open-circuit voltage of the lithium battery; is a correction function based on the state of health (SOH) of the battery.

7. A lithium battery SOC-SOH combined estimation device, characterized in that The device includes: A health factor extraction module, configured to model the lithium battery with a second-order RC equivalent circuit model, simulate the dynamic behavior of the lithium battery under different charge and discharge conditions by using the second-order RC equivalent circuit model, and extract a multi-dimensional battery health factor vector; A health state estimation module, configured to analyze the change trend of the multi-dimensional battery health factor vector by using a support vector machine regression model to obtain the SOH estimation value; A state of charge estimation module, configured to estimate the SOC based on the second-order RC equivalent circuit model by using the extended Kalman filter algorithm, and correct the SOC estimation process through the SOH estimation value to obtain the SOC estimation value; Wherein, the step of simulating the dynamic behavior of the lithium battery under different charge and discharge conditions by using the second-order RC equivalent circuit model and extracting the multi-dimensional battery health factor vector specifically includes: Simulate the voltage response of the lithium battery under different charge and discharge conditions by using the second-order RC equivalent circuit model, divide the charging voltage range during the charging process of the lithium battery into several voltage intervals, and count the interval charging time of each voltage interval; Use the sliding window method to obtain the average current in each voltage interval during the battery charging process of the lithium battery, and count the current change rate in the constant voltage stage; Extract the battery capacity increment curve from the charge and discharge data of the lithium battery during the charge and discharge process, and extract the peak parameter of the capacity increment curve from the battery capacity increment curve; wherein, the peak parameter of the capacity increment curve includes the maximum peak of the battery capacity increment curve, the time point corresponding to the maximum peak, and the second peak value; Use the numerical integration method to calculate the area of the battery capacity increment curve, and calculate the symmetry index of the capacity increment curve according to the area difference between the peak parameter of the capacity increment curve and the areas on the left and right sides of the battery capacity increment curve; Construct a multi-dimensional battery health factor vector based on the interval charging time, the average current value, the current change rate in the constant voltage stage, and the symmetry index of the capacity increment curve.

8. A computer device, characterized in that: It includes a processor and a memory. The processor is connected to the memory. The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the computer device executes the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: A computer program is stored in the computer-readable storage medium, and when the computer program is run, the method according to any one of claims 1 to 6 is implemented.

Images