Charge state estimation method, device, equipment and medium

Abstract

The invention provides a state of charge estimation method and device, equipment and a medium. The method comprises the following steps: acquiring a first actually measured current value of a battery pack under a preset time step length; the first actually-measured current value is input into an estimation model to estimate a first potential difference under a preset time step length, the estimation model is used for representing the relation between the potential difference and the current value of an equivalent circuit, the equivalent circuit is the equivalent circuit established for the battery pack, and the potential difference is the difference value between the open-circuit voltage value and the terminal voltage value of the equivalent circuit; calculating a first terminal voltage value of the battery pack under a preset time step according to the first potential difference; based on the first actually measured current value, establishing a state equation for estimating the state of charge of the battery pack; based on the equivalent circuit, establishing an observation equation for estimating the state of charge of the battery pack, the first terminal voltage value of the battery pack being an observation value; and based on the state equation and the observation equation, performing iteration by using a particle filter algorithm to obtain a state-of-charge estimation value of the battery pack.

Description

Technical Field

[0001] The present application relates to the field of energy storage technology, and in particular to a method, device, equipment and medium for estimating a state of charge. Background Art

[0002] Accurate estimation of battery state of charge is the core of battery management system, which is directly related to battery safety, performance optimization and life extension. Current state of charge estimation methods include traditional methods, model-based algorithms and intelligent algorithms, among which equivalent circuit-based estimation is particularly popular, but its accuracy is limited by the accurate identification of model parameters.

[0003] Parameter identification is divided into two categories: offline and online. Although offline identification has high accuracy, it is difficult to capture the time-varying characteristics of the battery. In online identification, although the least squares method can reflect dynamic characteristics, it faces the challenges of large computational complexity and insufficient accuracy; although the recursive least squares method can correct parameters, the data saturation problem can easily lead to increased deviations; the least squares method with a forgetting factor improves the convergence speed through weighted processing, but it performs poorly under constant current conditions and may even cause algorithm instability. Summary of the invention

[0004] In view of this, the purpose of the present application is to provide a method, device, equipment and medium for estimating the state of charge to overcome at least one of the above-mentioned defects.

[0005] In a first aspect, the present application provides a method for estimating a state of charge, the method comprising: obtaining a first measured current value of a battery pack at a preset time step; inputting the first measured current value into an estimation model to estimate a first potential difference at the preset time step, wherein the estimation model is used to characterize the relationship between the potential difference and the current value of an equivalent circuit, the equivalent circuit is an equivalent circuit established for the battery pack, and the potential difference is the difference between an open circuit voltage value and a terminal voltage value of the equivalent circuit; according to the first potential difference, calculating a first terminal voltage value of the battery pack at the preset time step; based on the first measured current value, establishing a state equation for estimating the state of charge of the battery pack; based on the equivalent circuit, establishing an observation equation for estimating the state of charge of the battery pack, wherein the first terminal voltage value of the battery pack is used as an observation value; based on the state equation and the observation equation, iterating using a particle filter algorithm to obtain an estimated value of the state of charge of the battery pack.

[0006] In one possible implementation, the estimation model is constructed in the following manner: establishing the Laplace equation of the equivalent circuit, wherein the Laplace equation is used to characterize the mathematical relationship between the potential difference and the current value of the equivalent circuit; obtaining a transfer function between the potential difference and the current value based on the Laplace equation; and representing the transfer function using the estimation model.

[0007] In one possible implementation, the estimation model is trained in the following manner: historical data of the battery pack is acquired, the historical data including second measured current values ​​of the battery pack at multiple preset time steps; based on the equivalent circuit of the battery pack, a second potential difference corresponding to the multiple second measured current values ​​is calculated; and the estimation model is trained using multiple training samples, each training sample including a second measured current value and a second potential difference at a preset time step.

[0008] In one possible implementation, the state equation is expressed using the following formula:

[0009] X(k+1)=AX(k)+BI(k)+W(k)

[0010] X(k)=[Cap k ,SOC k ] T

[0011]

[0012] Where X(k) is the vector state of the battery pack at time k, I(k) is the measured current value of the battery pack at time k, W(k) is the process noise, Cap k is the capacity of the battery pack at time k, SOC k is the state of charge of the battery pack at time k, and SOH is the ratio of the actual capacity to the nominal capacity.

[0013] In one possible implementation, the observation equation is expressed using the following formula:

[0014] Z(k)=U L (k)+V(k)

[0015] Among them, U L (k) is the terminal voltage of the battery pack at time k; U(k) is the potential difference of the battery pack at time k, and V(k) is the observation noise.

[0016] In one possible implementation, the step of iterating using a particle filter algorithm to obtain an estimated state of charge value of the battery pack includes: (A) generating a group of initial sampling particles according to a prior probability density function, wherein the initial state of charge of each initial sampling particle represents a state of charge value corresponding to an open circuit voltage value; (B) at the tth moment, calculating the likelihood function of the state of each sampling particle according to the estimated charge value of each sampling particle from the t-1th moment to the tth moment predicted by the state equation at the t-1th moment and the observed value obtained by the observation equation, and finally using the result of the likelihood function to update each sampling particle in the tth group of sampling particles. ; (C) resampling the tth group of sampled particles according to the weights to generate the t+1th group of sampled particles; (D) estimating the estimated state of charge of the battery pack at the tth moment according to the posterior probability density function, and predicting the estimated state of charge from the tth moment to the t+1th moment according to the state equation; (E) determining whether the tth moment has reached the preset time step; if the preset time step has not been reached, continuing the iteration to make t=t+1, and returning to execute step (B); if the preset time step has been reached, ending the iteration, and calculating the estimated state of charge of the battery pack according to the tth group of sampled particles.

[0017] In one possible implementation, the weight of each sampling particle in the tth group of sampling particles is calculated using the following formula:

[0018]

[0019] in, is the weight of the i-th particle at the t-1th moment, The state of the i-th particle at time t Next, observe The probability of is the observation value calculated by the observation equation at time t, is the probability distribution of state transition from time t-1 to time t, is used to generate the i-th particle at time t The importance density function, Z 1:t-1 is the set of all observations from time 1 to time t-1.

[0020] In a second aspect, the present application provides a state of charge estimation device, the device comprising: an acquisition module, used to acquire a first measured current value of a battery pack at a preset time step; a prediction module, used to input the first measured current value into an estimation model to estimate a first potential difference at the preset time step, wherein the estimation model is used to characterize the relationship between the potential difference and the current value of an equivalent circuit, the equivalent circuit is an equivalent circuit established for the battery pack, and the potential difference is the difference between an open circuit voltage value and a terminal voltage value of the equivalent circuit; a calculation module, used to calculate the first terminal voltage value of the battery pack at the preset time step according to the first potential difference; a first establishment module, used to establish a state equation for estimating the state of charge of the battery pack based on the first measured current value; a second establishment module, used to establish an observation equation for estimating the state of charge of the battery pack based on the equivalent circuit, wherein the first terminal voltage value of the battery pack is used as an observation value; an estimation module, based on the state equation and the observation equation, iterates using a particle filter algorithm to obtain an estimated value of the state of charge of the battery pack.

[0021] In a third aspect, the present application also provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of a method for estimating a state of charge as described above are performed.

[0022] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of a method for estimating a state of charge as described above are executed.

[0023] The present application provides a method, device, equipment and medium for estimating the state of charge, which utilizes a neural network to learn the functional relationship between the current and the potential difference of the equivalent circuit of a battery pack, replacing the traditional equivalent circuit parameter identification process and reducing the impact of parameter identification accuracy on the accuracy of state of charge estimation.

[0024] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 A flow chart of a method for estimating a state of charge provided in an embodiment of the present application;

[0027] Figure 2A A schematic diagram of the structure of an equivalent circuit provided in an embodiment of the present application;

[0028] Figure 2B A schematic diagram of the structure of the estimation model provided in the embodiment of the present application;

[0029] Figure 3 A flowchart showing the steps of constructing an estimation model provided by an exemplary embodiment of the present application;

[0030] Figure 4 A flowchart showing the steps of training an estimation model provided by an exemplary embodiment of the present application;

[0031] Figure 5 A flowchart showing steps of obtaining a state of charge estimation value of a battery pack by using a particle filtering algorithm provided by an exemplary embodiment of the present application;

[0032] Figure 6 A schematic diagram of the structure of a charge state estimation device provided in an embodiment of the present application;

[0033] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0035] The terms "a", "an", "the" and "said" are used in this specification to indicate the presence of one or more elements / components / etc.; the terms "including" and "having" are used to express an open-ended inclusion and mean that additional elements / components / etc. may exist in addition to the listed elements / components / etc.; the terms "first" and "second" etc. are used only as labels and are not intended to limit the quantity of their objects.

[0036] It should be understood that in the embodiments of the present application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" is merely a way to describe the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the previous and subsequent associated objects are in an "or" relationship. "Including A, B and / or C" means including any one, any two, or any three of A, B, and C.

[0037] It should be understood that in the embodiments of the present application, "B corresponding to A", "B corresponding to A", "A corresponds to B", or "B corresponds to A" means that B is associated with A, and B can be determined according to A. Determining B according to A does not mean determining B only according to A, and B can also be determined according to A and / or other information.

[0038] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0039] The battery state of charge (SOC) is one of the most important parameters of the battery management system, and its estimation accuracy is of great significance to ensuring battery safety, improving battery performance and extending battery life.

[0040] SOC estimation methods mainly include: traditional methods, model-based methods, and intelligent algorithms. Among them, the SOC estimation method based on equivalent circuits has been widely studied and applied, but the SOC estimation accuracy of this method depends on the accuracy of circuit model parameter identification. At present, parameter identification is divided into offline parameter identification and online parameter identification. Offline parameter identification is based on current and voltage data obtained from laboratory experiments, and the parameter identification is completed with the goal of highly fitting the predicted value with the measured value. This method has high parameter identification accuracy, but the parameter identification result cannot reflect the time-varying characteristics of the battery. The least squares algorithm, as a common online parameter identification method, can reflect the dynamic characteristics of the battery, but the algorithm has a large amount of calculation and poor parameter identification accuracy; the recursive least squares method can correct the parameters according to the observed data to obtain accurate parameter identification results, but with the increase in the number of acquisitions, data saturation will occur, resulting in large deviations in parameter estimation; the least squares algorithm with a forgetting factor weights the data at different times to avoid the rapid attenuation of the covariance matrix and the gain matrix and the loss of the ability to correct the estimated parameters, so that the algorithm can always maintain a faster convergence speed, but this algorithm cannot accurately identify the parameters of the constant current operating model, and may even cause the algorithm to diverge.

[0041] In response to at least one of the above-mentioned problems, the present application provides a method, device, equipment and medium for estimating the state of charge, which utilizes a neural network to learn the functional relationship between the current and the potential difference of the equivalent circuit of the battery pack, and accurately estimates the state of charge of the battery pack in real time.

[0042] See also Figure 1 , Figure 1 A flow chart of a method for estimating a state of charge provided in an embodiment of the present application.

[0043] So Figure 1 As shown in , the method for estimating the state of charge provided by the embodiment of the present application includes:

[0044] S101. Obtain a first measured current value of a battery pack at a preset time step.

[0045] In this step, a preset time interval is required during the data collection or processing process, and data measurement or processing will be performed once during this time interval. The current value of the battery pack obtained by measurement within the preset time step is the actual measurement result and is used for subsequent analysis and calculation.

[0046] S102: Inputting the first measured current value into an estimation model to estimate a first potential difference at a preset time step.

[0047] In this step, the estimation model may be a neural network model, which is used to characterize the relationship between the potential difference and the current value of the equivalent circuit.

[0048] like Figure 2A The figure shows a second-order RC equivalent circuit provided in an embodiment of the present application, which is an equivalent circuit established for a battery pack, and the potential difference is the difference between the open circuit voltage value and the terminal voltage value of the equivalent circuit.

[0049] Combine the following Figure 3 A preferred example of constructing an estimation model provided in an embodiment of the present application is introduced. It should be understood that the estimation model can also be constructed by various other existing methods, and the present application does not limit this.

[0050] Figure 3 A flowchart showing the steps of constructing an estimation model provided by an exemplary embodiment of the present application.

[0051] S201, establish the Laplace equation of the equivalent circuit,

[0052] In this step, the Laplace equation is used to characterize the mathematical relationship between the potential difference and the current value of the equivalent circuit. The Laplace equation of the second-order equivalent circuit is:

[0053]

[0054] Where: U L is the battery terminal voltage; U oc I is the battery open circuit voltage, which is a function of the battery state of charge (SOC) and has nothing to do with the current load current; m is the battery current; R is the ohmic internal resistance; R 1 and R 2 is the polarization internal resistance; C 1 and C 2 is a polarized capacitor.

[0055] S202. According to the Laplace equation, a transfer function between the potential difference and the current value is obtained.

[0056] In this step, the transfer function relationship between the open circuit voltage, the terminal voltage potential difference and the load current is obtained according to formula (1):

[0057]

[0058] Where: G(s) is the transfer function; τ 1 =R 1 C 1 ; τ 2 =R 2 C 2 .

[0059]

[0060] Where: T is the sampling period; z is the discretization operator.

[0061] Based on the bilinear transformation, the mathematical relationship between the battery voltage, current and model parameters in the z domain is established, and the discrete mathematical formulas between the battery voltage, current and model parameters are obtained as follows:

[0062] U(k)=U oc (k)-U L (k) = a 1 U(k-1)+a 2 U(k-2)+a 3 I m (k)+

[0063] a 4 I m (k-1)+a 5 I m (k-2)(4)

[0064] Where: U(k) is the potential difference between the output open circuit voltage and the terminal voltage; a 1 , a 2 , a 3 , a 4 , a 5 are the parameters to be identified.

[0065] S203: Express the transfer function using an estimation model.

[0066] In the embodiment of the present application, the estimation model may be a GRU (Gated Recurrent Unit) model. Figure 2B The structure of the GRU network provided by the embodiment of the present application is shown. The process of establishing the GRU network is described below.

[0067] Establish a GRU network to learn the functional relationship between the input sequence and the output sequence of the battery, where the input sequence is the current value and the output sequence is the potential difference. In GRU, each time step includes a hidden state h t and update gate r t , the update gate controls the information obtained from the prior state, and the hidden state is used to capture the long-term dependencies in the sequence. At each time step t, the GRU model predicts the potential difference as follows:

[0068] 1) Reset gate calculation: According to the reset gate r t Determines whether the model ignores the previous hidden state h t-1 Depends only on the current input x t .

[0069] r t =δ(W r [x t ,h t-1 ]) (5)

[0070] 2) Update gate calculation: According to the update gate z t Determine the new hidden state h of the model t Use the previous state h t-1 degree.

[0071]

[0072] 3) Calculate candidate hidden states: Candidate hidden states Using the current input x t and reset gate r t Estimate the next possible hidden state.

[0073]

[0074] 4) Calculate the hidden state: The final hidden state is weighted between the candidate hidden state and the previous hidden state, where the weight is determined by the update gate z t control.

[0075]

[0076] Where: x t 、h t-1 、h t 、r t 、z t , They are the current input vector, the output variable at the previous moment, the output variable at the current moment, the reset gate state, the update gate state, and the current candidate set state. t , W t , W are update gate, reset gate, candidate set and x respectively. t and h t-1 The weight value of the matrix product. δ and tanh represent the sigmoid function and tanh function respectively, [] represents vector connection, and * represents matrix product.

[0077] Combine the following Figure 4 Hereinafter, the process of training the estimation model provided in the embodiment of the present application will be introduced.

[0078] Figure 4 A flowchart showing the steps of training an estimation model provided by an exemplary embodiment of the present application.

[0079] S301. Obtain historical data of the battery pack.

[0080] In this step, the historical data includes the second measured current value of the battery pack at multiple preset time steps. This step is the process of data collection, the purpose of which is to obtain the measured current value of the battery pack at multiple different time points (i.e., preset time steps). These historical data are the basis for subsequent training of the estimation model. In addition, historical data is not limited to current values, but can also include other parameters related to battery performance such as temperature and voltage. In the embodiments of this application, the main focus is on current values.

[0081] S302. Calculate a second potential difference corresponding to a plurality of second measured current values ​​based on an equivalent circuit of the battery pack.

[0082] In this step, the potential difference corresponding to each measured current value is calculated using the equivalent circuit of the battery pack. By inputting the measured current value into the equivalent circuit, the potential difference between the corresponding open circuit voltage and the terminal voltage can be calculated, and these potential difference values ​​will be used as target values ​​when training the estimation model.

[0083] S303: Train the estimation model using multiple training samples.

[0084] In this step, the estimation model is trained using the historical data collected and processed previously (including measured current values ​​and potential differences). Each training sample contains a measured current value and a corresponding potential difference at a preset time step. By continuously and iteratively adjusting the parameters of the estimation model, the model can gradually learn the mapping relationship from current value to potential difference.

[0085] In one embodiment, in order to train a GRU-based neural network model, richer input features can be selected to improve the prediction accuracy of the model. Specifically, in addition to the current value at the current time step (time t), the temperature at time t, the open circuit voltage and terminal voltage potential difference at time t-1 and time t-2, and the corresponding current value are also considered. These input features together constitute the input vector of the neural network.

[0086] The output of GRU is the potential difference between the open circuit voltage and the terminal voltage at time t. By using a large amount of battery charging and discharging data to train this neural network, the complex laws of battery performance over time can be learned. Once the GRU network training is completed, new operating data (including current, temperature, and possible historical data) can be input into the model to predict or estimate the voltage response of the battery under load current.

[0087] This embodiment takes into account the time dependency and nonlinear characteristics of battery performance and improves the prediction capability of the model by introducing more historical information. At the same time, the GRU network, as a model that is good at processing sequence data, can effectively capture the changing trend of battery performance over time.

[0088] After completing the GRU network training, the current corresponding to the working condition is input into the neural network model to obtain the voltage response of the battery under the load current, where different working conditions may include but are not limited to constant current charging and discharging and variable working condition charging and discharging.

[0089] return Figure 1 , S103, calculating a first terminal voltage value of the battery pack at a preset time step according to the first potential difference.

[0090] In this step, the first potential difference (i.e., the difference between the open circuit voltage and the terminal voltage) previously obtained by the estimation model is used to calculate the terminal voltage value of the battery pack at the current preset time step. The terminal voltage is the voltage across the two ends of the battery pack under actual working conditions (i.e., when current passes through).

[0091] S104: Establishing a state equation for estimating the state of charge of the battery pack based on the first measured current value.

[0092] In this step, the particle filter algorithm SOC estimation state equation can be calculated by the ampere-hour integration method, and the calculation formula is:

[0093]

[0094] In the formula, SOC(t) and SOC(t 0 ) represent t 0 The SOC value at the moment and any moment; Q is the actual capacity of the battery; η is the Coulomb efficiency, usually 1; I(t) is the current at moment t, where the charging current is defined as negative and the discharging current is defined as positive.

[0095] In addition, the present application takes into account the effect of battery aging on battery capacity, wherein the SOC calculation formula taking into account the effect of battery aging on battery capacity is:

[0096]

[0097] Where: SOH is the ratio of actual capacity to nominal capacity. Discretizing the above formula, the equation for calculating SOC is:

[0098]

[0099] Since the battery capacity is a slowly varying function, it can be considered that the capacity at time k is consistent with the capacity at time k-1, that is:

[0100] Cap k =Cap k-1 (12)

[0101] Combining the battery capacity and SOC discrete equations, the state space equation is established. Assume that the state vector at time k is X(k) = [Cap k ,SOCk ] T , then the state equation for joint estimation of SOC and capacity is:

[0102] X(k+1)=AX(k)+BI(k)+W(k) (13)

[0103] Where: X(k) is the quantity to be estimated; W(k) is the process noise, and its covariance matrix is ​​Q. The state equation corresponding to the state vector X(k) is:

[0104]

[0105] Among them, the coefficient matrix is:

[0106]

[0107] S105. Based on the equivalent circuit, establish an observation equation for estimating the state of charge of the battery pack.

[0108] In this step, in order to more accurately estimate the SOC of the battery pack, in addition to establishing a state equation to describe the change of SOC over time, it is also necessary to establish an observation equation to estimate the SOC based on observable variables (such as terminal voltage). In this embodiment of the present application, the first terminal voltage value of the battery pack is used as the observation value.

[0109] In an embodiment of the present application, the open circuit voltage can be measured by leaving the battery to stand after being fully charged, and then discharging it under specific conditions, and measuring the open circuit voltage during each standing period. Specifically, at room temperature, the battery is fully charged, left to stand for 1 hour, and then discharged under 1 / 3C constant current conditions, and each time 5% of the capacity is released, it is left to stand for 1 hour. The voltage in this state is used as the open circuit voltage, and then the discharge is continued until the battery is discharged.

[0110] Since there is a nonlinear relationship between the open circuit voltage and SOC, it is necessary to fit this relationship through experimental data. In the embodiment of the present application, a 6th order polynomial is selected to fit the relationship between the open circuit voltage and SOC. This is because the polynomial function can flexibly approximate various nonlinear relationships, and the 6th order polynomial provides sufficient fitting accuracy while avoiding excessive complexity. The 6th order polynomial curve fitting obtains the relationship between the open circuit voltage and SOC as follows:

[0111] U OC =-6.4934*SOC 4 +14.4384*SOC 3 -10.366*SOC 2 +3.5739*SOC+3.1518 (17)

[0112] Where: U OCis the open circuit voltage. According to the potential difference between the open circuit voltage and the output of the neural network, the battery terminal voltage calculation formula is:

[0113] U L (k) = U OC (k)-U(k) (18)

[0114] Where: U L (k) is the battery terminal voltage; U(k) is the potential difference between the open circuit voltage and the terminal voltage. The observation equation established based on the circuit model is:

[0115] Z(k)=U L (k)+V(k) (19)Where V(k) is the observation noise, its covariance matrix is ​​R.

[0116] return Figure 1 , S106, based on the state equation and the observation equation, use the particle filter algorithm to iterate to obtain the estimated value of the state of charge of the battery pack.

[0117] In the battery management system, in order to accurately estimate the SOC, the state equation and the observation equation are usually combined, and an algorithm called particle filter (PF) is used for iterative calculation. Particle filtering is a nonlinear filtering technology based on Bayesian estimation and Monte Carlo method, which is particularly suitable for systems that process nonlinear and non-Gaussian noise.

[0118] Combine the following Figure 5 A preferred example of using a particle filter algorithm to iterate to obtain a state of charge estimate of a battery pack is introduced in an embodiment of the present application. It should be understood that the state of charge estimate of the battery pack can also be obtained through various other existing algorithms, and the present application does not impose any limitation on this.

[0119] Figure 5 A flowchart showing the steps of obtaining a state of charge estimation value of a battery pack by using a particle filtering algorithm provided by an exemplary embodiment of the present application.

[0120] S401. Generate a group of initial sampling particles according to a priori probability density function.

[0121] Specifically, the SOC estimation based on the state equation and observation equation is divided into two stages: prediction and update. The prediction stage generates a set of sampling particles according to the prior probability density function. Each particle contains all the information representing the SOC and capacity changes. These SOC values ​​can be mapped to the corresponding open circuit voltage values ​​through the known OCV-SOC relationship curve (i.e., OCV curve). Each sample is obtained through the system model to obtain the prior sample at time t, and the state of particle i is obtained as:

[0122]

[0123] Where: (·) (i) represents the i-th particle; i is the particle number; is the prior probability density function.

[0124] S402. At the tth moment, the likelihood function of the state of each sampling particle is calculated based on the charge estimation value of each sampling particle from the t-1th moment to the tth moment predicted by the state equation at the t-1th moment and the observation value obtained by the observation equation, and finally the result of the likelihood function is used to update the weight of each sampling particle in the tth group of sampling particles.

[0125] Specifically, in the prediction stage, a set of sampled particles are generated according to the prior probability density function. These particles contain all the information representing the SOC and capacity changes, and the prior samples at time t are obtained through the state equation. The state of each particle is predicted based on its historical state and the current system dynamics.

[0126] In the update phase, after receiving the measurement result Z(t), it is necessary to evaluate the consistency of each prior sample (i.e., each particle) with the new measurement result, that is, whether the SOC value represented by each particle may lead to the observation of such a terminal voltage, and obtain the particle weight as:

[0127]

[0128] Where: is the likelihood function of the particle filter algorithm.

[0129] Specifically, the weight calculation formula (21) takes into account three main factors: the weight of the previous moment (reflecting the importance of the particle in the previous iteration), the likelihood function It evaluates the possibility of observing the current terminal voltage under the SOC value represented by the current particle, as well as the importance density function It is simplified here to the prior probability density function (as shown in formula 22) because it is assumed in this example that the prior probability density function is a good importance density function. Therefore, the formula of the importance density function is:

[0130]

[0131] After calculating the weights, they need to be normalized to ensure that the sum of the weights of all particles is 1. The normalized weight calculation formula is:

[0132]

[0133] The posterior probability density function is approximated by the weighted sum of Dirac functions, namely:

[0134]

[0135] In the formula: is the posterior probability density function, X t represents the state of charge of the battery pack at time t, represents the sum of all items from i = 1 to i = Ns, where Ns is the total number of particles in a group. represents the weight of the i-th particle at the t-th time, is the i-th particle at time k, and δ(·) is the Dirac function. The posterior probability density function is a discrete probability distribution in which each particle represents a possible SOC value and its weight reflects the likelihood of that value.

[0136] In the embodiment of the present application, the likelihood function is selected as a Gaussian function, which reflects the closeness between the particle state (by estimating the terminal voltage) and the measured terminal voltage. The larger the Gaussian function, the closer the SOC value represented by the particle is to the true value, so the weight of the particle should also be larger, and its calculation formula is:

[0137]

[0138] Where: U L (t) is the estimated terminal voltage; U(t) is the measured terminal voltage.

[0139] S403: resample the tth group of sampling particles according to the weights to generate the t+1th group of sampling particles.

[0140] In this step, resampling is required according to the normalized weights to remove particles with lower weights and copy particles with higher weights, thereby obtaining a set of sampled particles (t+1) that better represent the true state of the system.

[0141] S404 , estimating the estimated state of charge of the battery pack at the tth moment according to the posterior probability density function, and predicting the estimated state of charge from the tth moment to the t+1th moment according to the state equation.

[0142] In this step, in order to estimate the SOC value at time t, the expected value of the posterior probability density function can be used as an estimate. In particle filtering, since the weighted sum of Dirac functions is used to approximate the posterior probability density function, the estimated value is usually obtained by calculating the weighted sum of all particle SOC values.

[0143] After obtaining the estimated SOC value at time t, the state equation can be used to predict the SOC value from time t to time t+1. The state equation usually describes how the SOC changes over time and how this change is affected by external factors such as battery current and temperature. Therefore, the estimated SOC value at time t can be used as the input of the state equation and the predicted SOC value at time t+1 can be calculated. This predicted value will be used as a priori estimate for the particle filter algorithm in the next iteration.

[0144] S405: Determine whether the tth moment has reached a preset time step.

[0145] In this step, the algorithm checks whether the current number of iterations (or time steps) has reached the preset time step. This time step can be a fixed number of iterations or a setting based on actual time (such as seconds, minutes, etc.).

[0146] If the preset time step is not reached, the iteration is continued to make t=t+1, and the process returns to step S402.

[0147] In this step, if the preset time step is not reached, the algorithm will perform the following operations: increase the time step counter t by 1 (i.e., t=t+1) to indicate the entry into the next iteration cycle, and return to execute step S402, that is, re-predict and update the process. The algorithm will use the state equation to predict the new SOC value based on the new time step (i.e., the new value of t) and the current system state (such as battery current, voltage, etc.), and use the observation equation and the new measurement data to update the particle weight.

[0148] If the preset time step is reached, step S406 is executed to end the iteration, and the estimated value of the state of charge of the battery pack is calculated based on the tth group of sampled particles.

[0149] In this step, if the preset time step is reached, it means that the algorithm has completed enough iterations and the iteration process can be terminated. At this time, the algorithm will execute step S406, that is, calculate the estimated value of the state of charge of the battery pack based on the last group (i.e., the tth group) of sampled particles. This estimated value is obtained based on the weighted sum of all particles, where particles with higher weights contribute more to the final estimated value.

[0150] Based on the same inventive concept, an embodiment of the present application also provides a state of charge estimation device corresponding to a state of charge estimation method. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned state of charge estimation method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0151] See also Figure 6 , Figure 6A schematic diagram of the structure of a charge state estimation device provided in an embodiment of the present application is shown in FIG. Figure 6 As shown in , the charge state estimation device 500 includes:

[0152] An acquisition module 501 is used to acquire a first measured current value of the battery pack at a preset time step;

[0153] A prediction module 502 is used to input the first measured current value into an estimation model to estimate a first potential difference at a preset time step, wherein the estimation model is used to characterize the relationship between the potential difference and the current value of the equivalent circuit, the equivalent circuit is an equivalent circuit established for the battery pack, and the potential difference is the difference between the open circuit voltage value and the terminal voltage value of the equivalent circuit;

[0154] A calculation module 503, configured to calculate a first terminal voltage value of the battery pack at a preset time step according to the first potential difference;

[0155] A first establishing module 504, configured to establish a state equation for estimating the state of charge of the battery pack based on the first measured current value;

[0156] A second establishing module 505 is used to establish an observation equation for estimating the state of charge of the battery pack based on the equivalent circuit, wherein the voltage value of the first terminal of the battery pack is used as the observation value;

[0157] The estimation module 506 uses a particle filter algorithm to iterate based on the state equation and the observation equation to obtain an estimated value of the state of charge of the battery pack.

[0158] See also Figure 7 , Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 7 As shown in , the electronic device 600 includes a processor 610 , a memory 620 and a bus 630 .

[0159] The memory 620 stores machine-readable instructions executable by the processor 610. When the electronic device 600 is running, the processor 610 communicates with the memory 620 via the bus 630. When the machine-readable instructions are executed by the processor 610, the above-mentioned Figure 1 The steps of a method for estimating a state of charge in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0160] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1As well as the steps of a method for estimating a state of charge in the method embodiment shown in FIG. 2 , the specific implementation method can be referred to the method embodiment, which will not be repeated here.

[0161] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0162] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0163] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0164] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0165] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, or in other words, the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0166] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for estimating a state of charge, characterized in that: The method comprises: Obtaining a first measured current value of the battery pack at a preset time step; Inputting the first measured current value into an estimation model to estimate a first potential difference at the preset time step, wherein the estimation model is used to characterize the relationship between the potential difference and the current value of an equivalent circuit, the equivalent circuit is an equivalent circuit established for the battery pack, and the potential difference is the difference between an open circuit voltage value and a terminal voltage value of the equivalent circuit; Calculating a first terminal voltage value of the battery pack at the preset time step according to the first potential difference; Based on the first measured current value, establishing a state equation for estimating the state of charge of the battery pack; Based on the equivalent circuit, establishing an observation equation for estimating the state of charge of the battery pack, wherein the voltage value of the first terminal of the battery pack is used as an observation value; Based on the state equation and the observation equation, a particle filter algorithm is used to iterate to obtain an estimated value of the state of charge of the battery pack.

2. The method according to claim 1, characterized in that The estimation model is constructed in the following way: Establishing a Laplace equation of the equivalent circuit, wherein the Laplace equation is used to characterize a mathematical relationship between a potential difference and a current value of the equivalent circuit; According to the Laplace equation, a transfer function between the potential difference and the current value is obtained; The transfer function is represented by an estimation model.

3. The method according to claim 2, characterized in that The estimation model is trained in the following way: Acquire historical data of the battery pack, the historical data comprising second measured current values ​​of the battery pack at a plurality of preset time steps; Calculating a second potential difference corresponding to a plurality of second measured current values ​​based on an equivalent circuit of the battery pack; The estimation model is trained using a plurality of training samples, each training sample comprising a second measured current value and a second potential difference at a preset time step.

4. The method according to claim 1, characterized in that The state equation is expressed using the following formula: X(k+1)=AX(k)+BI(k)+W(k) X(k)=[Cap k ,SOC k ] T Wherein, X(k) is the vector state of the battery pack at the kth moment, I(k) is the measured current value of the battery pack at the kth moment, W(k) is the process noise, Cap k is the capacity of the battery pack at the kth moment, SOC k is the state of charge of the battery pack at the kth moment, and SOH is the ratio of the actual capacity to the nominal capacity.

5. The method according to claim 1, characterized in that The observation equation is expressed using the following formula: Z(k)=U L (k)+V(k) Where Z(k) is the observed value of the battery pack at the kth moment, U L (k) is the terminal voltage of the battery pack at the kth moment, and V(k) is the observed noise.

6. The method according to claim 1, characterized in that The step of iterating using a particle filter algorithm to obtain an estimated value of the state of charge of the battery pack includes: (A) generating a set of initial sampling particles according to a priori probability density function, wherein the initial state of charge of each initial sampling particle represents a state of charge value corresponding to an open circuit voltage value; (B) at time t, according to the charge estimation value of each sampled particle in the tth group of sampled particles predicted by the state equation at time t-1 from time t-1 to time t and the observation value obtained by the observation equation, the likelihood function of the state of each sampled particle is calculated, so as to update the weight of each sampled particle using the result of the likelihood function; (C) resampling the tth group of sampling particles according to the weights to generate the t+1th group of sampling particles; (D) estimating the estimated state of charge of the battery pack at the tth moment according to the posterior probability density function, and predicting the estimated state of charge from the tth moment to the t+1th moment according to the state equation; (E) determining whether the t-th moment has reached the preset time step; If the preset time step is not reached, continue iterating to make t=t+1, and return to execute step (B); If the preset time step is reached, the iteration is terminated, and the estimated value of the state of charge of the battery pack is calculated based on the tth group of sampled particles.

7. The method according to claim 6, characterized in that The weight of each sampling particle in the tth group of sampling particles is calculated using the following formula: in, is the weight of the i-th particle at the t-1th moment, is the state of the i-th particle at time t Next, observe The probability of is the observation value calculated by the observation equation at time t, is the probability distribution of state transition from time t-1 to time t, is used to generate the i-th particle at time t The importance density function, Z 1:t-1 is the set of all observations from time 1 to time t-1.

8. A device for estimating a state of charge, characterized in that: The device comprises: An acquisition module, used to acquire a first measured current value of the battery pack at a preset time step; a prediction module, configured to input the first measured current value into an estimation model to estimate a first potential difference at the preset time step, wherein the estimation model is used to characterize the relationship between the potential difference and the current value of an equivalent circuit, the equivalent circuit is an equivalent circuit established for the battery pack, and the potential difference is a difference between an open circuit voltage value and a terminal voltage value of the equivalent circuit; a calculation module, configured to calculate a first terminal voltage value of the battery pack at the preset time step according to the first potential difference; A first establishing module, used to establish a state equation for estimating the state of charge of the battery pack based on the first measured current value; A second establishing module, configured to establish an observation equation for estimating the state of charge of the battery pack based on the equivalent circuit, wherein the voltage value of the first terminal of the battery pack is used as an observation value; The estimation module uses a particle filter algorithm to iterate based on the state equation and the observation equation to obtain an estimated value of the state of charge of the battery pack.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of any method as claimed in claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are executed.

Images