A method, apparatus, device, and medium for state estimation of energy storage batteries.

Abstract

This invention relates to the field of battery management technology and discloses a method, apparatus, device, and medium for state estimation of energy storage batteries. It uses an extended Kalman filter to initially estimate the state of the energy storage battery, obtaining the initial state value, terminal voltage error, and Kalman gain at the current moment. Using the current terminal voltage error and Kalman gain as input, an adaptive spiral flying sparrow search algorithm is used for error optimization, outputting an error compensation amount to compensate the initial state value. The compensated state value is then fed back to the extended Kalman filter, which includes recalibration and exit mechanisms, to estimate the state of the energy storage battery at the next moment. Based on the estimation result, the error optimization step and the state feedback step are repeated until a preset convergence condition is met, outputting the state estimation result of the energy storage battery. Through algorithmic collaboration and dynamic adjustment mechanisms, the accuracy and robustness of the energy storage battery state estimation are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to a method, apparatus, device, and medium for state estimation of energy storage batteries. Background Technology

[0002] Currently, energy storage systems, with their millisecond-level bidirectional power regulation capabilities and flexible operation characteristics, have become an important technical means to improve the absorption level of distributed power sources, improve power quality, and support the power supply of important loads during the distribution network fault recovery process. The mainstream energy storage method of current energy storage systems is battery energy storage. Therefore, battery state estimation (state of charge (SOC) and state of health (SOH)) is crucial for capturing the dynamic changes of battery energy storage systems in actual operation.

[0003] In existing technologies, data-driven methods are typically used to estimate the state of energy storage batteries. However, this method requires a large amount of high-quality data, while in practical applications, the data may be incomplete or contain noise, which in turn affects the reliability and accuracy of the estimation results.

[0004] Therefore, improving the reliability and accuracy of energy storage battery state estimation has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and medium for state estimation of energy storage batteries, addressing the problem of how to improve the reliability and accuracy of state estimation for energy storage batteries.

[0006] To address the aforementioned technical problems, the first aspect of this invention provides a state estimation method for energy storage batteries, comprising:

[0007] The identification parameters of the energy storage battery are obtained, and the state of the energy storage battery is initially estimated by an extended Kalman filter based on the identification parameters to obtain the initial state value, terminal voltage error and Kalman gain at the current moment.

[0008] Using the current terminal voltage error and Kalman gain as input, an error optimization algorithm is used to optimize the error and output the error compensation amount to compensate the initial state value, thereby obtaining the compensated estimated state value; the preset optimization algorithm is an adaptive spiral flight sparrow search algorithm.

[0009] The compensated estimated state value is fed back to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment. Based on the estimation result, the error optimization step and the state feedback step are repeated until the preset convergence condition is reached, and the state estimation result of the energy storage battery is output. The improved extended Kalman filter is configured to include an extended Kalman filter with recalibration and exit mechanisms.

[0010] As one preferred embodiment, the step of initially estimating the state of the energy storage battery using an extended Kalman filter based on the identification parameters to obtain the initial state value, terminal voltage error, and Kalman gain at the current moment includes:

[0011] The health state and state of charge of the energy storage battery are used as state variables, and state equations and observation equations are constructed based on the state variables and the identification parameters.

[0012] The state of the energy storage battery is predicted by the state equation and the observation equation, and the predicted state value and its corresponding prediction covariance matrix are obtained.

[0013] The state of the energy storage battery at the current moment is updated based on the predicted state value and its corresponding predicted covariance matrix to obtain the initial state value, terminal voltage error and Kalman gain at the current moment.

[0014] As one preferred embodiment, the step of taking the current terminal voltage error and Kalman gain as input, performing error optimization through a preset optimization algorithm, and outputting the error compensation amount includes:

[0015] A state estimation model for the energy storage battery is constructed based on the number of charge-discharge cycles, and an objective function is constructed with the minimum mean square error of the state estimation model as the optimization objective.

[0016] Using the current terminal voltage error and Kalman gain as inputs, the adaptive spiral flying sparrow search algorithm is used to optimize the initial state value based on the objective function, and the error compensation amount is output.

[0017] As one preferred embodiment, the adaptive spiral flight sparrow search algorithm is used to optimize the initial state value and output the error compensation amount, including:

[0018] The population is initialized using a tent chaotic mapping, and the solution space corresponding to the error compensation amount is divided into multiple concentric search rings according to a spiral topology, so that each ring region is assigned to each individual after being encoded in polar coordinates; each individual in the population represents a set of error compensation amounts.

[0019] The objective function is used as a fitness function to quantify the fitness values ​​of each individual after position update to determine the discoverer and follower;

[0020] The location of the discoverer is updated by introducing adaptive weights of the residuals of the extended Kalman filter or the residuals of the improved extended Kalman filter and the Levy flight mechanism to generate new candidate solutions in the sub-region where the discoverer is located, and the fitness of the new candidate solutions is calculated to update the discoverer.

[0021] The followers are updated using a variable spiral search strategy that incorporates time variables and the number of charge-discharge cycles of the energy storage battery;

[0022] The fitness values ​​of the updated discoverer and the updated follower are calculated based on the fitness function. Individuals with fitness values ​​lower than the preset fitness threshold are eliminated, and the population is re-initialized at the ring boundary. Individuals with fitness values ​​higher than the preset fitness threshold are retained to generate a new population.

[0023] The discovery update step and the follower update step are executed iteratively according to the new population until the preset number of iterations is reached, and the position of the individual with the highest fitness is selected as the error compensation amount for output.

[0024] As one preferred embodiment, the adaptive weights are expressed by the following formula:

[0025]

[0026] In the formula, w and w0 are the adaptive weight and the basic weight, respectively; For adjustment coefficients; RMS r M represents the residual of the extended Kalman filter or the residual of the improved extended Kalman filter; M is the maximum number of iterations.

[0027] As one preferred embodiment, the variable spiral search strategy is expressed by the following formula:

[0028]

[0029]

[0030] In the formula, x i (t), x worst (t) represents the position and worst position of the i-th individual in the t-th iteration, respectively; r is the dynamically adjusted spiral parameter; Step size control parameter; Q and L are random factors; A is the random matrix; n is the total population size; r0 is the initial dynamic adjustment spiral parameter; T is time; T max The maximum value of time; D cycle D cycle,max These represent the number of charge-discharge cycles and their maximum values ​​for the energy storage battery.

[0031] As one preferred embodiment, feeding the compensated estimated state value back into the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment includes:

[0032] In response to the compensated estimated state value and triggering recalibration, the compensated estimated state is used as the updated initial state value, so that the improved extended Kalman filter quantizes the target covariance matrix corresponding to the updated initial state value through the extended Kalman filter.

[0033] The target covariance matrix is ​​compared with the preset matrix increment. When the target covariance matrix satisfies the preset matrix increment, the exit mechanism is triggered, and the noise covariance is reset to the initial value, waiting for the next iteration. When the target covariance matrix does not satisfy the preset matrix increment, the state of the energy storage battery at the next moment is estimated based on the updated initial state value and its corresponding target covariance matrix.

[0034] A second aspect of the present invention provides a state estimation device for an energy storage battery, comprising:

[0035] The estimation module is used to obtain the identification parameters of the energy storage battery, and based on the identification parameters, to preliminarily estimate the state of the energy storage battery through an extended Kalman filter, so as to obtain the initial state value, terminal voltage error and Kalman gain at the current moment.

[0036] The compensation module is used to take the current terminal voltage error and Kalman gain as input, perform error optimization through a preset optimization algorithm, and output the error compensation amount to compensate the initial state value to obtain the compensated estimated state value; the preset optimization algorithm is an adaptive spiral flight sparrow search algorithm;

[0037] The output module is used to feed back the compensated estimated state value to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment, and repeat the error optimization step and state feedback step based on the estimation result until the preset convergence condition is reached, and output the state estimation result of the energy storage battery; the improved extended Kalman filter is configured as an extended Kalman filter that includes a recalibration and exit mechanism.

[0038] A third aspect of the present invention provides an electronic device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the state estimation method for an energy storage battery as described above.

[0039] A fourth aspect of the present invention provides a computer-readable storage medium comprising a stored computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the state estimation method for an energy storage battery as described above.

[0040] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following:

[0041] (1) The extended Kalman filter is used to make a preliminary estimate of the state of charge and health of the energy storage battery, which can effectively handle the noise of the nonlinear system; the estimation error is optimized and compensated by the adaptive spiral flying sparrow search algorithm, which further reduces the deviation of the terminal voltage error and Kalman gain, and significantly improves the state estimation accuracy.

[0042] (2) A recalibration mechanism is introduced in the extended Kalman filter, which can dynamically adjust the covariance matrix to adapt to external disturbances such as battery aging and temperature changes. When the variance of the predicted state increases significantly, the update is paused and the previous state estimate is retained to prevent the estimation divergence caused by abnormal data.

[0043] (3) The adaptive spiral flight sparrow search algorithm can dynamically adjust the error compensation amount through chaotic mapping, adaptive weight and spiral search strategy to adapt to the state changes of the battery under different charging and discharging rates, temperatures and environments; the compensated state value is fed back to the improved extended Kalman filter to form a closed-loop optimization process of "estimation-compensation-re-estimation" to ensure the real-time performance and accuracy of state estimation.

[0044] (4) By combining the improved extended Kalman filter and the adaptive spiral flight sparrow search algorithm, and introducing a recalibration and exit mechanism, high-precision estimation and dynamic optimization of the state of the energy storage battery are achieved. Attached Figure Description

[0045] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart of a state estimation method for energy storage batteries provided in a certain embodiment of the present invention;

[0047] Figure 2 This is a structural diagram of a state estimation device for an energy storage battery provided in a certain embodiment of the present invention;

[0048] Figure 3 This is a structural diagram of an electronic device provided in a certain embodiment of the present invention;

[0049] Figure label:

[0050] Among them, 10 is the estimation module; 20 is the compensation module; 30 is the output module; 5000 is the electronic device; 5001 is the processor; 5002 is the bus; 5003 is the memory; and 5004 is the transceiver. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings and examples. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0052] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0053] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0054] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is merely for describing specific embodiments and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0055] In one embodiment, such as Figure 1As shown, the first aspect of the present invention provides a state estimation method for an energy storage battery, comprising:

[0056] S1. Obtain the identification parameters of the energy storage battery, and based on the identification parameters, preliminarily estimate the state of the energy storage battery through an extended Kalman filter to obtain the initial state value, terminal voltage error and Kalman gain at the current moment; wherein, the identification parameters include the capacity, internal resistance, voltage and current of the energy storage battery.

[0057] In one embodiment, the step of initially estimating the state of the energy storage battery using an extended Kalman filter based on the identification parameters to obtain the initial state value, terminal voltage error, and Kalman gain at the current moment includes:

[0058] The health state and state of charge of the energy storage battery are used as state variables, and state equations and observation equations are constructed based on the state variables and the identification parameters.

[0059] The state of the energy storage battery is predicted by the state equation and the observation equation, and the predicted state value and its corresponding prediction covariance matrix are obtained.

[0060] The state of the energy storage battery at the current moment is updated based on the predicted state value and its corresponding predicted covariance matrix to obtain the initial state value, terminal voltage error and Kalman gain at the current moment.

[0061] Specifically, this invention uses an extended Kalman filter to estimate the state of the energy storage battery based on the acquired identification parameters:

[0062] This invention uses the State of Health (SOH) and State of Charge (SOC) of an energy storage battery as state variables, and sets initial state estimates (based on the battery's initial conditions or prior knowledge) and an initial covariance matrix (to represent the uncertainty of the initial state estimate), and sets the process noise covariance matrix Q. k To represent the uncertainty in the state transition process, the observation noise covariance matrix R is set. k This represents the uncertainty in the observation process. Then, based on the identified parameters and state variables, a state equation is constructed to describe the changes in battery state variables, namely, the state of health and the state of charge (SOH), over time. For SOH, its changes are usually related to factors such as the number of charge-discharge cycles and temperature, and the state equation can be constructed using empirical formulas or data-driven methods. For SOC, its changes are usually related to the battery's charge-discharge current, and the state equation can be constructed based on the ampere-hour integral method or an equivalent circuit model. Therefore, the state equation can be expressed as:

[0063]

[0064] In the formula, x k Let f be the actual state of the energy storage battery at time step k (i.e., the current SOH and SOC); f is the state transition function; u k For input vectors (such as charging / discharging current, open-circuit voltage, etc. in the identification parameters); w k This is process noise, and it follows a covariance of Q. k The zero-mean multivariate Gaussian distribution.

[0065] The observation equations are used to describe the relationship between the battery's state variables and measurable variables (such as terminal voltage and current). For SOC (State of Charge), the observation equations are constructed based on the relationship between the battery's open-circuit voltage and SOC. For SOH (State of Harshness), the observation equations involve the battery's internal resistance or other aging-related measurable parameters. Therefore, the state equations can be expressed as:

[0066]

[0067] In the formula, z k h(x) represents the measurement value of the energy storage battery at time step k (i.e., the observation vector at the current moment); k ) is the mapping function from the state vector to the observation vector; v k Measurement noise that follows a zero-mean multivariate Gaussian distribution.

[0068] Next, based on the state equation and the current state estimate (and its corresponding covariance matrix), the state value is predicted. Simultaneously, based on the state equation and the covariance matrix of the process noise, the covariance matrix corresponding to the previous state estimate is predicted. This helps assess the uncertainty of the predicted state. The prediction of the covariance matrix is ​​performed using the following formula:

[0069]

[0070] In the formula, P k|k-1 To predict the covariance matrix; F k Let f be the state transition matrix (a linearized approximation of f), and let f be the partial derivatives of the state transitions with respect to the state.

[0071] Then, using the current observations and the observation equation, combined with the results of the prediction step, the Kalman gain is calculated:

[0072]

[0073] In the formula, K k H is the Kalman gain; k Let be the observation matrix, and let represent the partial derivatives of the observations with respect to the state.

[0074] The state estimate at the current moment is quantized using Kalman gain and observations, and is expressed by the following formula:

[0075]

[0076] In the formula, To predict the state at time k based on information from time (k-1), assume an estimation error. Follow the variance Given a zero-mean multivariate Gaussian distribution, the first step of each iteration is prediction, denoted as . .

[0077] And update the covariance matrix using the following formula:

[0078]

[0079] In the formula, I is the identity matrix.

[0080] Finally, the updated state estimates are... The output is the initial state value at the current moment, and during the update process, the terminal voltage error (i.e., the observed value) also needs to be calculated. Compared with the predicted value (Differences between them).

[0081] This invention uses both the health state and the state of charge (SOC) as state variables, which can more comprehensively reflect the performance state of the battery. By constructing state equations and observation equations, the dynamic behavior of the battery can be described more accurately, thereby improving the accuracy of state estimation. The Kalman filter algorithm itself has the ability to suppress noise. Through iterative updates of the state equations and observation equations, the estimation error can be gradually reduced. Furthermore, by introducing the health state as a state variable, the impact of battery aging on SOC estimation can be further considered, enhancing the robustness of the algorithm.

[0082] S2. Using the current terminal voltage error and Kalman gain as input, perform error optimization through a preset optimization algorithm, and output the error compensation amount to compensate the initial state value to obtain the compensated estimated state value; the preset optimization algorithm is the Adaptive Spiral Flying Sparrow Search Algorithm (ASFSSA).

[0083] In one embodiment, the step of using the current terminal voltage error and Kalman gain as inputs to perform error optimization through a preset optimization algorithm and outputting an error compensation amount includes:

[0084] A state estimation model for the energy storage battery is constructed based on the number of charge-discharge cycles, and an objective function is constructed with the minimum mean square error of the state estimation model as the optimization objective.

[0085] Using the current terminal voltage error and Kalman gain as inputs, the adaptive spiral flying sparrow search algorithm is used to optimize the initial state value based on the objective function, and the error compensation amount is output.

[0086] Specifically, this invention constructs a state estimation model for energy storage batteries based on the number of charge-discharge cycles. , It is expressed by the following formula:

[0087]

[0088] In the formula, This represents the maximum capacity of the energy storage battery. The initial state of charge of the energy storage battery is denoted as I(t); I(t) is the current at time t. This is a correction term for the SOC estimate; These are state-of-charge parameters; This refers to the number of cycles of the energy storage battery; A parameter that depends on health status; The parameter can be determined from historical data.

[0089] The objective function is constructed with the goal of minimizing the mean square error of the state estimation model, and is expressed by the following formula:

[0090]

[0091] In the formula, , The estimation errors for SOC and SOH are respectively determined by their mean square errors.

[0092] The state of health (SOH) of an energy storage battery typically exhibits a non-linear relationship with its cycle count. This is mainly manifested in the fact that as the cycle count increases, the battery's capacity decay and internal resistance increase more rapidly. Therefore, this invention introduces the cycle count as a parameter, which can more accurately reflect the long-term use of the battery and capture the dynamic trend of its health status.

[0093] The SOH estimation model typically includes parameters such as terminal voltage, current, and temperature. However, these parameters mainly reflect the instantaneous state of the battery, while the number of cycles is a cumulative indicator that can supplement the shortcomings of instantaneous parameters in describing long-term performance degradation, thereby improving the robustness and accuracy of the SOH estimation model.

[0094] The calculated terminal voltage error and Kalman gain are used as inputs to the adaptive spiral flight sparrow search algorithm. The objective function is used as the fitness function to calculate the fitness function value based on the input. By iteratively updating the sparrow's position and speed, the algorithm finds the sparrow that minimizes the fitness function value and uses it as the error compensation amount to compensate for the initial state value obtained based on the extended Kalman filter, thus obtaining the compensated estimated state value. The process is shown in the following equation:

[0095] SOC opt = SOC EKF +ΔSOC

[0096] SOH opt = SOH EKF +ΔSOH

[0097] In the formula, SOC opt SOH opt To compensate for the estimated state value; SOC EKF SOH EKF ΔSOC and ΔSOH are the initial state values; ΔSOC and ΔSOH are the error compensation values.

[0098] This invention constructs a state estimation model based on the number of charge-discharge cycles and optimizes it based on the objective function that minimizes the mean square error, thus more accurately reflecting the actual state of the battery. The application of the adaptive spiral flying sparrow search algorithm enables more efficient searching for the optimal solution in the solution space, further reducing the error in state estimation. After introducing the adaptive spiral flying sparrow search algorithm for optimization, and using the number of cycles as an additional input dimension, the model is provided with richer feature information, improving its ability to capture the nonlinear variation law of SOH. Furthermore, constructing the state estimation model based on the number of charge-discharge cycles takes into account the battery aging process, enabling the model to adapt to changes in battery state under different cycle counts.

[0099] In one embodiment, the step of using the adaptive spiral flight sparrow search algorithm to optimize the initial state value and outputting an error compensation amount includes:

[0100] The population is initialized using a tent chaotic mapping, and the solution space corresponding to the error compensation amount is divided into multiple concentric search rings according to a spiral topology, so that each ring region is assigned to each individual after being encoded in polar coordinates; each individual in the population represents a set of error compensation amounts.

[0101] The objective function is used as a fitness function to quantify the fitness values ​​of each individual after position update to determine the discoverer and follower;

[0102] The location of the discoverer is updated by introducing adaptive weights of the residuals of the extended Kalman filter or the residuals of the improved extended Kalman filter and the Levy flight mechanism to generate new candidate solutions in the sub-region where the discoverer is located, and the fitness of the new candidate solutions is calculated to update the discoverer.

[0103] The followers are updated using a variable spiral search strategy that incorporates time variables and the number of charge-discharge cycles of the energy storage battery;

[0104] The fitness values ​​of the updated discoverer and the updated follower are calculated based on the fitness function. Individuals with fitness values ​​lower than the preset fitness threshold are eliminated, and the population is re-initialized at the ring boundary. Individuals with fitness values ​​higher than the preset fitness threshold are retained to generate a new population.

[0105] The discovery update step and the follower update step are executed iteratively according to the new population until the preset number of iterations is reached, and the position of the individual with the highest fitness is selected as the error compensation amount for output.

[0106] Specifically, to improve the randomness of population initialization, this invention uses tent chaotic mapping to initialize the population, making the individual distribution more uniform and enhancing population diversity: an initial value is randomly generated, and a chaotic sequence is iteratively generated using the tent chaotic mapping formula. This chaotic sequence is then mapped into the search space of the error compensation quantity to obtain the initial population; wherein,

[0107] The formula for the chaotic mapping of the tent is as follows:

[0108]

[0109] In the formula, b i A chaotic sequence for a tent chaotic mapping; Let L be the parameter vector for the j-th individual, including the state of charge error compensation and the health state error compensation; L and U are the upper and lower bounds of the model parameters. The j-th value in the chaotic sequence generated by the tent chaotic mapping is used to randomly initialize the parameter vector within the upper and lower bounds of the model parameters. This mapping guarantees a uniform distribution of individuals in the solution space, improving the quality of population initialization.

[0110] Set algorithm parameters, such as population size, discoverer ratio, initial dynamic adjustment spiral parameter, maximum number of iterations, etc., and then use the objective function as the fitness function to quantify the fitness value of each individual. The smaller the fitness value, the better the error compensation for that individual. Based on the quantization results and the set algorithm parameters, the sparrows in the population are assigned roles: select several individuals with better fitness values ​​as discoverers based on the discoverer ratio, randomly select several individuals as watchers based on the watcher ratio, and the remaining individuals as followers.

[0111] The discoverer position is updated using adaptive weights introduced from the residuals of an extended Kalman filter or an improved extended Kalman filter, along with the Levy flight mechanism; wherein the adaptive weights are expressed by the following formula:

[0112]

[0113]

[0114]

[0115] In the formula, w and w0 are the adaptive weight and the basic weight, respectively; For adjustment coefficients; RMS r is the residual of the extended Kalman filter or the residual of the improved extended Kalman filter; M is the maximum number of iterations; W is the size of the sliding window.

[0116] Therefore, the location of the discoverer is updated by introducing adaptive weights:

[0117]

[0118] In the formula, x i (t) represents the position of the i-th individual in the t-th iteration.

[0119] It should be noted that the adaptive weights update the discoverer's location using the residuals of the extended Kalman filter only in the initial update; subsequent updates use the residuals of the improved extended Kalman filter. The current residuals of the Kalman filter (i.e., the root mean square of the deviation between the observed and predicted values ​​in the state estimate within the window) are used to dynamically adjust the search step size for individual locations. This allows for a wider search range to accelerate convergence when the residuals are large, and a more refined search to improve accuracy when the residuals are small.

[0120] The formula for Levy's flight mechanism is as follows:

[0121]

[0122] In the formula, The step size is a control parameter, dynamically adjusted by adaptive weights w to control the search step size and direction; x p Let λ be the current globally optimal individual position; Levy(λ) is a random path following the Levy distribution. Through the Levy flight mechanism, individuals can jump within the solution space with a random step size combining long and short steps, enhancing their ability to escape local optima.

[0123] After updating the discoverer's location based on adaptive weights and the Levy flight mechanism, new candidate solutions can be generated in the sub-region where the discoverer is located. The fitness of the new candidate solutions is calculated according to the fitness function and compared with the fitness value of the original discoverer. The individual with the better fitness value is selected as the updated discoverer.

[0124] Next, a variable spiral search strategy incorporating time variables and the number of charge-discharge cycles of the energy storage battery is adopted to update the follower's position; the update of the follower's position by the variable spiral search strategy is expressed by the following formula:

[0125]

[0126]

[0127] In the formula, x worst (t) represents the worst position of the i-th individual in the t-th iteration; r is the dynamically adjusted spiral parameter; Q and L are random factors; A is the random matrix; n is the total population size; r0 is the initial dynamically adjusted spiral parameter; T is the time; T max The maximum value of time; D cycle D cycle,max These represent the number of charge-discharge cycles of the energy storage battery and its maximum value. This invention employs a variable spiral search strategy that incorporates time variables and the number of charge-discharge cycles of the energy storage battery. By dynamically adjusting the spiral parameters, the search path becomes more diverse. This strategy results in a larger spiral search amplitude in the early stages of iteration and when the battery is new, which is beneficial for global search; while in the later stages of iteration and when the battery is aging, the spiral search amplitude gradually decreases, which is beneficial for local search and convergence.

[0128] Then, the fitness values ​​of the updated discoverer and the updated follower are calculated based on the fitness function. Individuals with fitness values ​​lower than the preset fitness threshold are eliminated according to the calculation results. The fitness values ​​are re-initialized at the ring boundary, and individuals with fitness values ​​higher than the preset fitness threshold are retained to generate a new population. Finally, the discovery update step and the follower update step are iteratively executed based on the new population until the preset number of iterations is reached, and the position of the individual with the highest fitness is selected as the error compensation amount for output.

[0129] This invention employs a tent chaotic mapping to initialize the population, increasing its diversity and ergodicity, which helps avoid the algorithm getting trapped in local optima and thus improves optimization accuracy. The introduction of adaptive weights and the Levy flight mechanism allows the algorithm to dynamically adjust the search step size and direction during the search process, further increasing the likelihood of finding the global optimum. The variable spiral search strategy, by introducing time variables and the number of charge-discharge cycles of the energy storage battery, makes the search process more flexible and efficient, accelerating the algorithm's convergence speed. The introduction of a residual factor as an adaptive adjustment factor achieves a dynamic trade-off between estimation accuracy (error minimization) and estimation confidence (covariance trace minimization), avoiding overfitting or divergence problems caused by a single objective. The introduction of online closed-loop optimization of the noise covariance matrix suppresses drastic fluctuations in the Kalman gain. This scheme comprehensively considers the number of charge-discharge cycles of the energy storage battery and the objective function, enabling the algorithm to adapt to changes in battery state under different operating conditions and enhancing its robustness. By optimizing the initial state value and outputting the error compensation amount, the accuracy of energy storage battery state estimation can be significantly improved, providing a more reliable basis for battery management.

[0130] In another embodiment, the symbiotic biogeography optimization-Levi flight hybrid algorithm can also be used as a preset optimization algorithm for error optimization to obtain the error compensation amount, which specifically includes:

[0131] Initialize the number of habitats, migration-in / migration rate calculation parameters, and Levy flight parameters. Randomly generate an initial population for each habitat using a set of error compensation values. Construct a fitness function using the current terminal voltage error and Kalman gain to quantify the fitness value of each habitat.

[0132] Based on the fitness values ​​and the initialized immigration and emigration rates, the immigration and emigration rates are quantified using calculation parameters to perform immigration and emigration operations on each of the habitats.

[0133] Based on the initialized Levi flight parameters, a Levi flight step size is generated to apply Levi flight perturbation to the migrated habitat;

[0134] Two habitats are randomly selected to exchange some features for mutually beneficial cooperation, and the fitness value of the new habitat is calculated. The habitat with the higher fitness value is retained to generate a new population.

[0135] Based on the new population, the migration in and out operations and Levi flight disturbances are repeatedly performed until the preset number of iterations is reached, and the habitat with the highest fitness value is selected as the error compensation output.

[0136] The symbiotic biogeographical optimization-Levie flight hybrid algorithm used in this invention for error optimization combines the species migration mechanism of biogeographical optimization with the long-distance jump characteristics of Levie flight to perform the optimization process, including:

[0137] First, the parameters for calculating habitat quantity and immigration / escape rates, such as baseline immigration / escape rate values ​​and adjustment factors, as well as Lévy flight parameters such as Lévy flight step size coefficient and Lévy flight index, are initialized. Then, a set of error compensation values ​​is randomly generated within the solution space as the initial population for each habitat, ensuring that each habitat represents a possible solution space region. Based on the current terminal voltage error and Kalman gain, a fitness function is constructed to quantify the fitness value of each habitat, i.e., the quality of the error compensation value corresponding to each habitat. The fitness function can be expressed by the following formula:

[0138]

[0139] In the formula, Fitness i X is the fitness value; i Let be the i-th habitat; λ be the mean immigration rate.

[0140] Based on the calculated fitness values ​​and initialized immigration and emigration rates, parameters are used to quantify the immigration and emigration rates. For each habitat, a trait is used to predict whether individuals migrate to other habitats based on the immigration rate (e.g., ΔSOC or ΔSOH), and a trait is used to predict whether individuals migrate to other habitats based on the emigration rate. Immigration operations move individuals from habitats with low fitness values ​​to habitats with high fitness values, while emigration operations move individuals from habitats with high fitness values ​​to habitats with low fitness values, thereby increasing population diversity.

[0141] The Lévy flight step size is generated based on the initialized Lévy flight parameters. The Lévy flight step size follows a Lévy distribution, which can be generated by the Mantegna algorithm. A Lévy flight perturbation is applied to the migrated habitat, i.e., a random search is performed around it to explore new solution space regions.

[0142] Two habitats are randomly selected and some of their characteristics (such as certain dimensions of the error compensation quantity) are exchanged to engage in mutually beneficial cooperation. Through this cooperation, new solutions can be generated and the fitness of the population may be improved. The fitness value of the newly generated habitat is calculated and compared with the fitness value of the original habitat. The habitat with the higher fitness value is retained to generate a new population.

[0143] Finally, based on the new population, the migration-in and migration-out operations and Levy flight disturbances are repeated until the preset number of iterations is reached. In each iteration, the population is updated according to the new habitat and fitness value. Finally, the error compensation amount corresponding to the habitat with the highest fitness value is selected as the final output. This error compensation amount can be used to correct the initial state value of the energy storage battery state estimation model and improve the accuracy of state estimation.

[0144] This invention balances exploration and exploitation by simulating the migration and emigration behaviors of species in an ecosystem; it introduces Lévy flight into the local search to enhance the ability to escape local optima; it improves the quality of solutions through mutually beneficial cooperation among species; and it significantly improves the solution efficiency of complex nonlinear optimization problems through species migration, Lévy flight perturbation, and symbiotic cooperation mechanisms. Furthermore, this scheme can adapt to different types of energy storage batteries and different operating states, demonstrating strong versatility and practicality.

[0145] S3. Feed the compensated estimated state value back to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment, and repeat the error optimization step and state feedback step based on the estimation result until the preset convergence condition is reached, and output the state estimation result of the energy storage battery; the improved extended Kalman filter is configured as an extended Kalman filter that includes a recalibration and exit mechanism.

[0146] In one embodiment, feeding the compensated estimated state value back into the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment includes:

[0147] In response to the compensated estimated state value and triggering recalibration, the compensated estimated state is used as the updated initial state value, so that the improved extended Kalman filter quantizes the target covariance matrix corresponding to the updated initial state value through the extended Kalman filter.

[0148] The target covariance matrix is ​​compared with the preset matrix increment. When the target covariance matrix satisfies the preset matrix increment, the exit mechanism is triggered, and the noise covariance is reset to the initial value, waiting for the next iteration. When the target covariance matrix does not satisfy the preset matrix increment, the state of the energy storage battery at the next moment is estimated based on the updated initial state value and its corresponding target covariance matrix.

[0149] Specifically, the "recalibration" step in the improved extended Kalman filter of this invention re-estimates the actual effect of the Kalman gain after the state update, allowing the state estimator to exit the update when the update is useless. That is, after compensating the initial state value, the compensated estimated state value can trigger recalibration. At this time, the compensated estimated state is used as the updated initial state value, so that the improved extended Kalman filter can quantify the target covariance matrix corresponding to the updated initial state value through the covariance matrix calculation method of the traditional extended Kalman filter.

[0150] This invention embeds the ASFSSA algorithm into the main loop of the designed improved extended Kalman filter to achieve online real-time optimization of noise parameters. Assume the process noise covariance matrix Q of the improved extended Kalman filter is... kThe observation noise covariance matrix R k Since all matrices are diagonal matrices, the optimization variables are:

[0151]

[0152] In the formula, q1, q2, q3, and r are the parameters to be optimized.

[0153] The stability of both prediction error and Kalman gain is considered:

[0154]

[0155] In the formula, denoted as , where b is the estimation error of the noise covariance moment; b is the balance coefficient, which is set to 0.1 in this invention; N is the total time step; Prediction Error is the prediction error term; Gain Stability is the gain stability term; and Trace is the Frobenius norm of the Kalman gain.

[0156] The updates to the process noise covariance matrix and the observation noise covariance matrix during each iteration can be expressed by the following formula:

[0157]

[0158] In the formula, These are the updated process noise covariance matrix and the updated observation noise covariance matrix, respectively.

[0159] By employing swarm intelligence search (ASFSSA), the local optima of traditional gradient methods are avoided. ASFSSA optimization is called once every E steps.

[0160]

[0161] In the formula, , These are the process noise covariance matrix and observation noise covariance matrix after ASFSSA optimization, respectively;

[0162] Finally, the updated covariance matrix is ​​used as the target covariance matrix in each iteration to estimate the state of the energy storage battery at the next moment.

[0163] Next, the obtained target covariance matrix is ​​compared with the preset matrix increment (the preset matrix increment can be set according to the actual needs of the system and the performance requirements of the filter, and is used to determine whether the target covariance matrix meets the requirements). When the target covariance matrix meets the preset matrix increment, it means that the current state estimation is accurate enough, triggering the exit mechanism, resetting the noise covariance to the initial value, and proceeding to the next iteration, that is, restarting the prediction-update step of the initial state value based on the initialized noise covariance. When the target covariance matrix does not meet the preset matrix increment, it means that the current state estimation still needs to be improved, so the state of the energy storage battery at the next moment is estimated based on the updated initial state value and its corresponding target covariance matrix.

[0164] The basic idea of ​​the improved extended Kalman filter used in this invention is not to reduce the updated state covariance matrix to the smallest possible value, but to re-estimate the system after the state update to more accurately estimate the actual effect of the Kalman gain. In fact, in the new framework, if the system is strongly nonlinear, the trace of the covariance matrix after recalibration may increase, and the predicted state may have a large variance. This phenomenon indicates that "recalibration" cannot make the estimation more accurate and should be withdrawn. To address this situation, the improved extended Kalman filter adds an "exit" step to the framework, allowing the state estimator to exit the update process when updating is no longer useful.

[0165] This invention, by compensating for estimated state values ​​and triggering recalibration, can more accurately reflect the actual state of the energy storage battery, reducing estimation deviations caused by initial state errors. Using the compensated estimated state as the updated initial state value helps improve the convergence and stability of the extended Kalman filter during the iteration process, increasing the accuracy of state estimation. When the target covariance matrix satisfies the preset matrix increment, an exit mechanism is triggered and the noise covariance is reset to its initial value, which can prevent the noise covariance from accumulating errors during the iteration process and maintain the stability of the filter's performance. This scheme can dynamically adjust the filter parameters and estimation process according to the actual state of the energy storage battery and environmental changes, improving the filter's adaptability and robustness.

[0166] After estimating the state of the energy storage battery at the next moment, the error optimization step and the state feedback step are repeatedly executed based on the estimation result until the preset convergence condition is reached, that is, when the preset number of iterations is reached, the final state estimation value is output as the state estimation result of the energy storage battery. Alternatively, the state estimation value of the next moment obtained in each iteration can be output as the state estimation result of the energy storage battery to analyze the state of the energy storage battery.

[0167] This application addresses the issue of improving the reliability and accuracy of energy storage battery state estimation by designing a state estimation method for energy storage batteries. The method involves acquiring the identification parameters of the energy storage battery and, based on these parameters, using an extended Kalman filter to initially estimate the battery's state, obtaining the initial state value, terminal voltage error, and Kalman gain at the current moment. Using the current terminal voltage error and Kalman gain as input, an error optimization algorithm is applied to optimize the error, outputting an error compensation amount to compensate the initial state value, resulting in a compensated estimated state value. The preset optimization algorithm is an adaptive spiral-flying sparrow search algorithm. The compensated estimated state value is fed back to an improved extended Kalman filter to estimate the state of the energy storage battery at the next moment. Based on the estimation result, the error optimization step and state feedback step are repeatedly executed until a preset convergence condition is met, outputting the state estimation result of the energy storage battery. The improved extended Kalman filter is configured with a technical solution including recalibration and exit mechanisms. Through algorithmic collaboration and dynamic adjustment mechanisms, the accuracy and robustness of energy storage battery state estimation are significantly improved.

[0168] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.

[0169] In another embodiment, such as Figure 2 As shown, a second aspect of the present invention provides a state estimation device for an energy storage battery, comprising:

[0170] The estimation module 10 is used to obtain the identification parameters of the energy storage battery, and based on the identification parameters, to preliminarily estimate the state of the energy storage battery through an extended Kalman filter, so as to obtain the initial state value, terminal voltage error and Kalman gain at the current moment.

[0171] The compensation module 20 is used to take the current terminal voltage error and Kalman gain as input, perform error optimization through a preset optimization algorithm, and output the error compensation amount to compensate the initial state value to obtain the compensated estimated state value; the preset optimization algorithm is an adaptive spiral flight sparrow search algorithm;

[0172] The output module 30 is used to feed back the compensated estimated state value to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment, and repeat the error optimization step and state feedback step based on the estimation result until the preset convergence condition is reached, and output the state estimation result of the energy storage battery; the improved extended Kalman filter is configured as an extended Kalman filter that includes a recalibration and exit mechanism.

[0173] It should be noted that the various modules in the aforementioned state estimation device for energy storage batteries can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module. For specific limitations regarding the state estimation device for energy storage batteries, please refer to the limitations regarding the state estimation method for energy storage batteries described above; both have the same function and role, and will not be repeated here.

[0174] A third aspect of the present invention provides an electronic device comprising:

[0175] Processor, memory, and bus;

[0176] The bus is used to connect the processor and the memory;

[0177] The memory is used to store operation instructions;

[0178] The processor is configured to execute instructions by invoking the operation instructions, causing the processor to perform operations corresponding to a state estimation method for an energy storage battery as shown in the first aspect of this application.

[0179] In one alternative embodiment, an electronic device is provided, such as Figure 3 As shown, Figure 3 The illustrated electronic device 5000 includes a processor 5001 and a memory 5003. The processor 5001 and the memory 5003 are connected, for example, via a bus 5002. Optionally, the electronic device 5000 may also include a transceiver 5004. It should be noted that in practical applications, the transceiver 5004 is not limited to one type, and the structure of this electronic device 5000 does not constitute a limitation on the embodiments of this application.

[0180] Processor 5001 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 5001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0181] Bus 5002 may include a path for transmitting information between the aforementioned components. Bus 5002 may be a PCI bus or an EISA bus, etc. Bus 5002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0182] The memory 5003 may be a ROM or other type of static storage device capable of storing static information and instructions, RAM or other type of dynamic storage device capable of storing information and instructions, or it may be an EEPROM, CD-ROM or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0183] The memory 5003 is used to store application code that executes the scheme of this application, and its execution is controlled by the processor 5001. The processor 5001 is used to execute the application code stored in the memory 5003 to implement the content shown in any of the foregoing method embodiments.

[0184] Among them, electronic devices include, but are not limited to: mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers.

[0185] The fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a state estimation method for an energy storage battery as shown in the first aspect of the present application.

[0186] Another embodiment of this application provides a computer-readable storage medium storing a computer program that, when run on a computer, enables the computer to execute the corresponding content in the aforementioned method embodiments.

[0187] Furthermore, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0188] In summary, this invention relates to the field of battery management technology and discloses a method, apparatus, device, and medium for state estimation of energy storage batteries. It uses an extended Kalman filter to initially estimate the state of the energy storage battery, obtaining the initial state value, terminal voltage error, and Kalman gain at the current moment. Using the current terminal voltage error and Kalman gain as input, an adaptive spiral flying sparrow search algorithm is used for error optimization, outputting an error compensation amount to compensate the initial state value. The compensated state value is then fed back to the extended Kalman filter, which includes recalibration and exit mechanisms, to estimate the state of the energy storage battery at the next moment. Based on the estimation result, the error optimization step and the state feedback step are repeatedly executed until a preset convergence condition is met, outputting the state estimation result of the energy storage battery. Through algorithmic collaboration and dynamic adjustment mechanisms, the accuracy and robustness of energy storage battery state estimation are significantly improved.

[0189] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts among the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; refer to the description of the method embodiments for relevant details. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0190] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A state estimation method for energy storage batteries, characterized in that, include: The identification parameters of the energy storage battery are obtained, and the state of the energy storage battery is initially estimated by an extended Kalman filter based on the identification parameters to obtain the initial state value, terminal voltage error and Kalman gain at the current moment. Using the current terminal voltage error and Kalman gain as input, an error optimization algorithm is used to optimize the error and output the error compensation amount to compensate the initial state value, thereby obtaining the compensated estimated state value; the preset optimization algorithm is an adaptive spiral flight sparrow search algorithm. The compensated estimated state value is fed back to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment. Based on the estimation result, the error optimization step and the state feedback step are repeated until the preset convergence condition is reached, and the state estimation result of the energy storage battery is output. The improved extended Kalman filter is configured to include an extended Kalman filter with recalibration and exit mechanisms; The step of initially estimating the state of the energy storage battery using an extended Kalman filter based on the identification parameters to obtain the initial state value, terminal voltage error, and Kalman gain at the current moment includes: The health state and state of charge of the energy storage battery are used as state variables, and state equations and observation equations are constructed based on the state variables and the identification parameters. The state of the energy storage battery is predicted by the state equation and the observation equation, and the predicted state value and its corresponding prediction covariance matrix are obtained. The state of the energy storage battery at the current moment is updated based on the predicted state value and its corresponding predicted covariance matrix to obtain the initial state value, terminal voltage error and Kalman gain at the current moment. The step of taking the current terminal voltage error and Kalman gain as input, performing error optimization through a preset optimization algorithm, and outputting an error compensation amount includes: A state estimation model for the energy storage battery is constructed based on the number of charge-discharge cycles, and an objective function is constructed with the minimum mean square error of the state estimation model as the optimization objective. Using the current terminal voltage error and Kalman gain as inputs, the adaptive spiral flying sparrow search algorithm is used to optimize the initial state value based on the objective function, and the error compensation amount is output.

2. The state estimation method for energy storage batteries according to claim 1, characterized in that, The adaptive spiral flight sparrow search algorithm is used to optimize the initial state value and output the error compensation amount, including: The population is initialized using a tent chaotic mapping, and the solution space corresponding to the error compensation amount is divided into multiple concentric search rings according to a spiral topology, so that each ring region is assigned to each individual after being encoded in polar coordinates; each individual in the population represents a set of error compensation amounts. The objective function is used as a fitness function to quantify the fitness values ​​of each individual after position update to determine the discoverer and follower; The location of the discoverer is updated by introducing adaptive weights of the residuals of the extended Kalman filter or the residuals of the improved extended Kalman filter and the Levy flight mechanism to generate new candidate solutions in the sub-region where the discoverer is located, and the fitness of the new candidate solutions is calculated to update the discoverer. The followers are updated using a variable spiral search strategy that incorporates time variables and the number of charge-discharge cycles of the energy storage battery; The fitness values ​​of the updated discoverer and the updated follower are calculated based on the fitness function. Individuals with fitness values ​​lower than the preset fitness threshold are eliminated, and the population is re-initialized at the ring boundary. Individuals with fitness values ​​higher than the preset fitness threshold are retained to generate a new population. The discovery update step and the follower update step are executed iteratively according to the new population until the preset number of iterations is reached, and the position of the individual with the highest fitness is selected as the error compensation amount for output.

3. The state estimation method for energy storage batteries according to claim 2, characterized in that, The adaptive weights are represented by the following formula: In the formula, w and w0 are the adaptive weight and the basic weight, respectively; For adjustment coefficients; RMS r t represents the residual of the extended Kalman filter or the residual of the improved extended Kalman filter; M is the maximum number of iterations; t is the current number of iterations.

4. The state estimation method for energy storage batteries according to claim 3, characterized in that, The variable spiral search strategy is expressed by the following formula: In the formula, x i (t), x worst (t) represents the position and worst position of the i-th individual in the t-th iteration, respectively; r is the dynamically adjusted spiral parameter; Step size control parameter; Q and L are random factors; A is the random matrix; n is the total population size; r0 is the initial dynamic adjustment spiral parameter; T is time; T max The maximum value of time; D cycle D cycle,max These represent the number of charge-discharge cycles and their maximum values ​​for the energy storage battery.

5. The state estimation method for energy storage batteries according to claim 1, characterized in that, The step of feeding the compensated estimated state value back to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment includes: In response to the compensated estimated state value and triggering recalibration, the compensated estimated state is used as the updated initial state value, so that the improved extended Kalman filter quantizes the target covariance matrix corresponding to the updated initial state value through the extended Kalman filter. The target covariance matrix is ​​compared with the preset matrix increment. When the target covariance matrix satisfies the preset matrix increment, the exit mechanism is triggered, and the noise covariance is reset to the initial value, waiting for the next iteration. When the target covariance matrix does not satisfy the preset matrix increment, the state of the energy storage battery at the next moment is estimated based on the updated initial state value and its corresponding target covariance matrix.

6. A state estimation device for an energy storage battery, characterized in that, include: The estimation module is used to obtain the identification parameters of the energy storage battery, and based on the identification parameters, to preliminarily estimate the state of the energy storage battery through an extended Kalman filter, so as to obtain the initial state value, terminal voltage error and Kalman gain at the current moment. The compensation module is used to take the current terminal voltage error and Kalman gain as input, perform error optimization through a preset optimization algorithm, and output the error compensation amount to compensate the initial state value to obtain the compensated estimated state value; the preset optimization algorithm is an adaptive spiral flight sparrow search algorithm; The output module is used to feed the compensated estimated state value back to the improved extended Kalman filter to estimate the state of the energy storage battery at the next moment, and repeat the error optimization step and state feedback step based on the estimation result until the preset convergence condition is reached, and output the state estimation result of the energy storage battery. The improved extended Kalman filter is configured to include an extended Kalman filter with recalibration and exit mechanisms; The step of initially estimating the state of the energy storage battery using an extended Kalman filter based on the identification parameters to obtain the initial state value, terminal voltage error, and Kalman gain at the current moment includes: The health state and state of charge of the energy storage battery are used as state variables, and state equations and observation equations are constructed based on the state variables and the identification parameters. The state of the energy storage battery is predicted by the state equation and the observation equation, and the predicted state value and its corresponding prediction covariance matrix are obtained. The state of the energy storage battery at the current moment is updated based on the predicted state value and its corresponding predicted covariance matrix to obtain the initial state value, terminal voltage error and Kalman gain at the current moment. The step of taking the current terminal voltage error and Kalman gain as input, performing error optimization through a preset optimization algorithm, and outputting an error compensation amount includes: A state estimation model for the energy storage battery is constructed based on the number of charge-discharge cycles, and an objective function is constructed with the minimum mean square error of the state estimation model as the optimization objective. Using the current terminal voltage error and Kalman gain as inputs, the adaptive spiral flying sparrow search algorithm is used to optimize the initial state value based on the objective function, and the error compensation amount is output.

7. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the state estimation method for an energy storage battery as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the state estimation method for an energy storage battery as described in any one of claims 1 to 5.

Images