Charge state prediction method and device, storage medium and electronic equipment

Abstract

The invention provides a state-of-charge prediction method and device, a storage medium and electronic equipment, and the method comprises the steps: obtaining a fractional order state space model corresponding to a battery equivalent circuit, wherein the state variable of the fractional order state space model comprises the state of charge; discretizing the fractional order state space model to obtain a discrete state space model; adjusting the noise of the discrete state space model to enable the noise of the discrete state space model to meet a Gaussian distribution condition, and obtaining a first discrete state space model; and performing state-of-charge prediction on the first discrete state space model based on an extended Kalman filtering algorithm. According to the method, the noise of the discrete state space model is adjusted, so that the noise of the discrete state space model meets the Gaussian distribution condition, a proper use condition can be provided for the Kalman filtering algorithm, the state of charge is predicted based on the extended Kalman filtering algorithm, and the prediction accuracy can be improved.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and more specifically to a method, apparatus, storage medium, and electronic device for predicting state of charge. Background Technology

[0002] With the widespread application of electric vehicles and energy storage systems, lithium iron phosphate batteries have attracted much attention due to their high safety and long cycle life. Among them, accurately estimating the state of charge (SOC) of lithium iron phosphate batteries is the key to ensuring their safe and efficient operation. The accuracy of the SOC estimation directly affects the battery's lifespan, performance, and economics.

[0003] Currently, SOC estimation based on the extended Kalman filter algorithm is the mainstream method. However, the accuracy of the current Kalman filter algorithm in estimating the state of charge depends on the accuracy of the battery model. Lithium iron phosphate batteries are complex electrochemical systems with complex nonlinear characteristics in their charging and discharging processes. This means that the relationship between SOC and battery current and voltage is complex and variable, which makes it difficult to further improve the accuracy of the battery model after reaching a bottleneck. Consequently, the accuracy of SOC estimation based on the extended Kalman filter algorithm cannot be guaranteed. Summary of the Invention

[0004] In view of this, embodiments of the present invention aim to provide a method, apparatus, storage medium, and electronic device for predicting the state of charge, in order to solve the problem that the accuracy of the prediction of the state of charge cannot be guaranteed in the prior art.

[0005] This invention provides a method for predicting the state of charge, the method comprising:

[0006] Obtain the fractional-order state-space model corresponding to the battery equivalent circuit, wherein the state variables of the fractional-order state-space model include the state of charge.

[0007] The fractional-order state-space model is discretized to obtain a discrete state-space model;

[0008] The noise of the discrete state space model is adjusted so that the noise of the discrete state space model satisfies the Gaussian distribution condition, thus obtaining the first discrete state space model.

[0009] The state of charge is predicted based on the extended Kalman filter algorithm of the first discrete state-space model, and the prediction result of the state of charge is obtained.

[0010] In one embodiment, the noise adjustment process for the discrete state-space model includes:

[0011] Noise estimates are added to the state variables of the discrete state-space model.

[0012] The noise estimate is updated based on the random walk algorithm, and the noise of the discrete state space model is generated based on the updated noise estimate.

[0013] In one embodiment, the noise of the first discrete state-space model includes process noise, and the prediction of the state of charge of the first discrete state-space model based on the extended Kalman filter algorithm includes:

[0014] Obtain the covariance matrix corresponding to the process noise and the actual parameters of the battery equivalent circuit, wherein the actual parameters characterize the current state of the battery equivalent circuit;

[0015] The element in the covariance matrix corresponding to the state of charge is determined as the target element, and the target element represents the reciprocal of the prediction confidence of the state of charge.

[0016] Based on the actual parameters, the target elements in the first discrete state space model are adjusted to obtain the second discrete state space model.

[0017] The state of charge of the second discrete state-space model is predicted based on the extended Kalman filter algorithm.

[0018] In one embodiment, the real-world parameters include the current, and adjusting the target elements in the first discrete state-space model according to the real-world parameters to obtain a second discrete state-space model includes:

[0019] If the magnitude of the current does not exceed the current threshold, then the current amplitude of the current is obtained;

[0020] The element value of the target element is determined based on the current amplitude, and the element value of the target element is positively correlated with the current amplitude.

[0021] If the magnitude of the current exceeds the current threshold, the element value of the target element remains unchanged.

[0022] In one embodiment, the real-world parameters include the current state of charge, and adjusting the target elements in the first discrete state-space model according to the real-world parameters to obtain a second discrete state-space model includes:

[0023] Based on the current state of charge, the element value of the target element is determined, wherein when the current state of charge is between the middle value and the upper value corresponding to the state of charge, the current state of charge is positively correlated with the element value of the target element; when the current state of charge is between the lower limit value and the middle value corresponding to the state of charge, the current state of charge is negatively correlated with the element value of the target element.

[0024] In one embodiment, the real-world parameters include a curve showing the relationship between the battery open-circuit voltage and the current state of charge. The step of adjusting the target elements in the first discrete state-space model based on the real-world parameters to obtain a second discrete state-space model includes:

[0025] Obtain the derivative corresponding to the relationship curve;

[0026] The element value of the target element is determined based on the derivative, and the derivative is positively correlated with the element value of the target element.

[0027] In one embodiment, the reality parameter includes the current information value, and adjusting the target element in the first discrete state-space model according to the reality parameter to obtain the second discrete state-space model includes:

[0028] Based on the current information value, the element value of the target element is determined, and the information value is positively correlated with the element value of the target element.

[0029] In one embodiment, the method further includes:

[0030] When the battery equivalent circuit is in a charging state, the current state of charge is obtained;

[0031] If the current state of charge is less than or equal to the state of charge threshold, then the state of charge of the battery equivalent circuit is predicted based on the ampere-hour integral algorithm to obtain the predicted state of charge.

[0032] If the current state of charge is greater than the state of charge threshold, then the step of performing state of charge prediction on the first discrete state-space model based on the extended Kalman filter algorithm is executed to obtain the prediction result of the state of charge.

[0033] In one embodiment, the battery equivalent circuit is a second-order R-CPE equivalent circuit containing fractional-order constant-phase elements. The fractional-order state-space model includes first-order state variables, second-order state variables, and third-order state variables. The first-order state variables characterize the voltage drop corresponding to the electrochemical polarization resistance in the second-order R-CPE equivalent circuit, the second-order state variables characterize the voltage drop corresponding to the concentration difference polarization resistance in the second-order R-CPE equivalent circuit, and the third-order state variables characterize the state of charge.

[0034] In one embodiment, the discrete state-space model includes fourth-order state variables, which represent noise estimates of the discrete state-space model.

[0035] Another aspect of the present invention provides a state of charge prediction device, the device comprising:

[0036] The model acquisition module is used to acquire the fractional-order state-space model corresponding to the battery equivalent circuit, wherein the state variables of the fractional-order state-space model include the state of charge.

[0037] The discrete processing module is used to discretize the fractional-order state-space model to obtain a discrete state-space model.

[0038] The noise adjustment module is used to adjust the noise of the discrete state space model so that the noise of the discrete state space model satisfies the Gaussian distribution condition, thereby obtaining the first discrete state space model.

[0039] The prediction module is used to predict the state of charge of the first discrete state-space model based on the extended Kalman filter algorithm, and obtain the prediction result of the state of charge.

[0040] In another aspect, the present invention provides a computer-readable storage medium having stored thereon computer-executable instructions which, when executed by a processor, implement the state of charge prediction method as described in any of the above embodiments.

[0041] In another aspect, the present invention provides an electronic device, the electronic device comprising:

[0042] processor;

[0043] Memory used to store the processor's executable instructions;

[0044] The processor is used to execute the state of charge prediction method described in any of the above embodiments.

[0045] Compared with related technologies, the state of charge prediction method provided by this invention has the following advantages:

[0046] The state of charge (SOC) prediction method provided by this invention includes: obtaining a fractional-order state-space model corresponding to the battery equivalent circuit, wherein the state variables of the fractional-order state-space model include the SOC; discretizing the fractional-order state-space model to obtain a discrete state-space model; adjusting the noise of the discrete state-space model to ensure that the noise satisfies the Gaussian distribution condition, thus obtaining a first discrete state-space model; and predicting the SOC based on the extended Kalman filter algorithm to obtain the predicted SOC result. Since using the Kalman filter algorithm for SOC prediction typically requires assuming that the noise of the discrete state-space model satisfies the Gaussian distribution condition, but the actual model error is difficult to satisfy this condition, this invention adjusts the noise of the discrete state-space model to ensure that the noise satisfies the Gaussian distribution condition. This provides suitable conditions for the Kalman filter algorithm, and using the extended Kalman filter algorithm to predict the SOC based on the first discrete state-space model improves the accuracy of SOC prediction. Attached Figure Description

[0047] Figure 1 The diagram shown is a flowchart of a state of charge prediction method provided in an embodiment of the present invention.

[0048] Figure 2 The diagram shown is a schematic diagram of the equivalent circuit of a battery provided in an embodiment of the present invention.

[0049] Figure 3 The diagram shown is a flowchart of a state of charge prediction method provided in another embodiment of the present invention.

[0050] Figure 4 The figure shown is a comparison curve of the state of charge provided in an embodiment of the present invention.

[0051] Figure 5 The figure shown is a comparison curve of the error of the state of charge provided in an embodiment of the present invention.

[0052] Figure 6 The figure shown is a comparison curve of the state of charge provided by another embodiment of the present invention.

[0053] Figure 7 The figure shown is a comparison curve of the error of the state of charge provided by another embodiment of the present invention.

[0054] Figure 8 The diagram shown is a schematic block diagram of a state of charge prediction device provided in an embodiment of the present invention.

[0055] Figure 9 The diagram shown is a block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] In view of the above problems, one embodiment of the present invention provides a method for predicting the state of charge, which can be executed by a computer device (e.g., a server or a user terminal). Figure 1 As shown, the state of charge prediction method may include:

[0058] 110. Obtain the fractional-order state-space model corresponding to the battery equivalent circuit. The state variables of the fractional-order state-space model include the state of charge.

[0059] In some implementations, such as Figure 2 As shown, the battery equivalent circuit can be a second-order R-CPE equivalent circuit containing fractional-order constant-phase elements. This second-order R-CPE equivalent circuit can include: a first resistor R0, a second resistor R1, and a third resistor R2, wherein the first resistor R0 is the ohmic internal resistance in the second-order R-CPE equivalent circuit, the second resistor R1 is the electrochemical polarization resistance in the second-order R-CPE equivalent circuit, and the third resistor R2 is the concentration difference polarization resistance in the second-order R-CPE equivalent circuit. Optionally, this battery equivalent circuit can be an equivalent circuit constructed for lithium iron phosphate batteries.

[0060] The second resistor R1 is connected in parallel with a first fractional-order constant-phase element CPE1 to form a first R-CPE branch, and the third resistor R2 is connected in parallel with a second fractional-order constant-phase element CPE2 to form a second R-CPE branch. The first fractional-order constant-phase element CPE1... PE1 The capacitance value can be expressed as C. CPE1 Represents the first fractional-order constant-phase element C. PE1 The order can be represented by α1, and the voltage drop of the first R-CPE branch can be represented by V1; the second fractional-order constant-phase element C PE2 The capacitance value can be expressed as C. CPE2 This indicates that the second fractional-order constant-phase element C PE2 The order can be represented by α2, and the voltage drop of the second R-CPE branch can be represented by V2.

[0061] The steady-state open-circuit voltage of the battery equivalent circuit can be used for V oc This indicates that the output voltage of the battery equivalent circuit can be V. o This indicates that the current in the equivalent circuit of the battery can be represented by I.

[0062] For example, according to Figure 2 The expression for the fractional-order state-space model constructed from the battery equivalent circuit shown can be as follows:

[0063]

[0064] y = [1 1 0]x + V oc (·)+R0(·)u (2)

[0065] Among them, expression (1) is the process equation of the fractional state-space model, and expression (2) is the observation equation (also referred to as the output equation) of the fractional state-space model.

[0066] Where u is the input of the fractional-order state-space model, and u = I, and y is the output of the fractional-order state-space model, and y = V. o Let x be the state variable of the fractional-order state-space model. The state variable x can include a first-order state variable x1, a second-order state variable x2, and a third-order state variable x3. The first-order state variable represents the voltage drop V1 corresponding to the electrochemical polarization resistance in the second-order R-CPE equivalent circuit, the second-order state variable represents the voltage drop V2 corresponding to the concentration difference polarization resistance in the second-order R-CPE equivalent circuit, and the third-order state variable represents the state of charge (SOC).

[0067] in, This represents the time differential operator d / dt. Describes the fractional differential operator d with respect to time. α / dt α η represents charge efficiency, C Q This indicates the battery capacity, expressed in Ah. The 3600 in the equation is a unit conversion factor used to convert battery capacity from ampere-hours (Ah) to coulombs (C). This conversion ensures that the unit of current u(k) is consistent with the unit of the rate of change of SOC, thus correctly reflecting the dynamic behavior of the battery in the state equation. The '(·)' indicates that the above variable is a function of multiple independent variables, which may optionally include SOC, temperature, state of charge / discharge, etc.

[0068] It is understood that, in this embodiment, the battery equivalent circuit can be a second-order R-CPE equivalent circuit, or an R-CPE equivalent circuit of other orders, such as a third-order or fourth-order R-CPE equivalent circuit. Correspondingly, for different orders of R-CPE equivalent circuits, the state variables in their fractional-order state-space models are also different. For example, for a third-order R-CPE equivalent circuit, its state variables can be... Wherein, V3 is the voltage drop of one R-CPE branch other than the first R-CPE branch and the second R-CPE branch.

[0069] 120. Discretize the fractional-order state-space model to obtain the discrete state-space model.

[0070] Following the example above, since the fractional-order state-space model is a continuous-time state-space model, it can be discretized to facilitate processing by computer devices.

[0071] Specifically, the sampling period T can be set according to the actual conditions. s (e.g., setting T) s =1s), assuming that the parameters change very slowly with respect to the state (i.e., the parameters are constant within one sampling period), and combining the discretization numerical algorithm of fractional calculus to discretize the fractional state-space model, we can obtain a discrete state-space model, the expression of which is as follows:

[0072]

[0073] y(k)=[1 1 0]x(k)+V oc (x3(k))+R0u(k) (4)

[0074] Among them, expression (3) is the process equation of the discrete state space model, and expression (4) is the observation equation of the discrete state space model.

[0075] Where, ω i,j Here, ω represents the coefficients used in the fractional calculus discretization calculation, i represents the index of the state variable, j represents the index of the time step during discretization, k represents the current time step, and L is the cutoff position in the fractional calculus discretization calculation. Optionally, the cutoff position L can be 80°. i,j The recursive calculation method can be shown below:

[0076]

[0077] When the cutoff position is +∞, the calculation result of fractional calculus is accurate. Any finite cutoff position will introduce cutoff error, which in turn will lead to steady-state error in the model.

[0078] To eliminate this steady-state error, we can apply this to all j≥1 ω i,j Perform uniform scaling to achieve the final result.

[0079] Therefore, this discrete state-space model can be simplified as follows:

[0080]

[0081] 130. Adjust the noise of the discrete state-space model so that the noise of the discrete state-space model satisfies the Gaussian distribution condition, and obtain the first discrete state-space model.

[0082] In some implementations, noise exists in the battery equivalent circuit due to inaccuracies in the battery equivalent circuit and the sampling of battery physical quantities. In this embodiment, corresponding noise can be introduced into the battery equivalent circuit to address the aforementioned inaccuracy problem. Specifically, a noise term can be added to the discrete state-space model to represent the noise of the discrete state-space model.

[0083] For example, by adding a noise term to the discrete state-space model shown in expressions (6) and (7), we can obtain the following expression for the discrete state-space model:

[0084]

[0085] y(k)=h(x(k), u(k))+v(k) (9)

[0086] The noise term can include w(k) and v(k), where w(k) is the process noise and v(k) is the observation noise.

[0087] In some implementations, the specific implementation of adjusting the noise of the discrete state space model in step 130 may include: amplifying the low-frequency part of the noise and weakening the high-frequency part of the noise to obtain an estimated value of the noise so that the noise of the discrete state space model satisfies the Gaussian distribution condition.

[0088] In some implementations, the specific implementation of adjusting the noise of the discrete state-space model in step 130 may include:

[0089] 131. Add noise estimates to the state variables of the discrete state-space model.

[0090] The discrete state-space model includes fourth-order state variables, which represent noise estimates of the discrete state-space model. For example, as described above... Add x4 to make the state variable Here, x4 is a fourth-order state variable, representing the noise estimate of the discrete state-space model.

[0091] As an example, adding a noise estimate x4 to the state variables of a discrete state-space model can update the discrete state-space model to the following expression:

[0092]

[0093] y(k)=[1 1 0 1]x(k)+V oc (x3(k))+R0u(k)+v(k) (11)

[0094] 132. Update the noise estimate based on the random walk algorithm, and generate the noise of the discrete state space model based on the updated noise estimate.

[0095] For example, the update equation for updating the noise estimate based on the random walk algorithm can be expressed as: x4(k+1)=x4(k)+w(k), where x4(k+1) is the updated noise estimate, w(k) is the noise process noise of the discrete state space model, which can usually be assumed to be Gaussian white noise, and x4(k) is the noise estimate before the update. The meaning of this update equation is that the noise estimate x4(k) at each time step k+1 is equal to its value at the previous time step k plus the process noise w(k) at the current time. x4 is modeled as a random walk process. Through this update method, x4 accumulates the historical noise w(k), thereby capturing the low-frequency characteristics of the noise. The random walk algorithm amplifies the low-frequency noise through the accumulation effect, while weakening the high-frequency noise, so that the noise of the discrete state space model satisfies the Gaussian distribution condition.

[0096] It can be understood that, since the above update equation gives the relationship between the updated noise estimate and the noise of the discrete state space model, the noise of the discrete state space model can be determined based on the updated noise estimate using this update equation.

[0097] As can be seen, in this embodiment, by adding a noise estimate x4 to the state variables of the discrete state-space model and updating the noise estimate x4 using a random walk algorithm, it is equivalent to introducing a shaping filter to reshape the noise characteristics of the discrete state-space model so that the noise characteristics of the discrete state-space model satisfy the Gaussian distribution condition, so as to facilitate subsequent prediction of the state of charge using the Kalman filter algorithm.

[0098] 140. Based on the extended Kalman filter algorithm, the state of charge of the first discrete state-space model is predicted, and the prediction results of the state of charge are obtained.

[0099] Following the example above, the expression for the first discrete state-space model can be found in expressions (10) and (11), and the noise estimate in the first discrete state-space model is updated using a random walk equation. For this noisy model, the covariance matrix corresponding to the process noise w in the first discrete state-space model can be represented by Q, and the covariance matrix corresponding to the observation noise v can be represented by R. Assuming the previous estimated time is k, the steps for each iteration of the fractional extended Kalman filter (EKF) algorithm to predict the state of charge of the first discrete state-space model are as follows:

[0100] S1. Using the information at time k, calculate the derivative F of the state equation with respect to the state (local linearization of the state equation):

[0101]

[0102] Where F(k) is the partial derivative of the state equation with respect to the state variable x(k), representing the state estimate at the current time k. Local linearization of the state equation under input u(k). Since the battery model is nonlinear, local linearization can approximate the nonlinear relationship as a linear relationship, thereby simplifying the subsequent prediction and update steps, so as to capture the dynamic behavior of the battery state as the input changes.

[0103] S2. Using the information at time k, predict the state at time k+1. and state covariance:

[0104]

[0105]

[0106] Among them, Ω j =[ω 1,j ω 2,j 0 0] T .

[0107] in, It predicts the state value at the next time k+1 based on the information at the current time k.

[0108] Wherein, P(k│k) is the covariance matrix of the predicted state, used to represent the uncertainty of the predicted state.

[0109] Where Q(k) is the covariance matrix of the process noise, used to represent the uncertainty of the model error.

[0110] S3. Using the information at time k+1, calculate the derivative H of the output equation with respect to the state (local linearization of the output equation):

[0111]

[0112] Where H(k+1) is the partial derivative of the observation equation with respect to the state variable x(k+1), representing the value of H(k+1) in the predicted state. Local linearization of the output equation under input u(k+1). Calculating the derivative H is to establish the relationship between the output (e.g., voltage) and state variables (e.g., SOC) for subsequent Kalman gain calculation.

[0113] S4. Calculate the Kalman gain K:

[0114]

[0115] Here, K(k+1) is the Kalman gain at time k+1, representing the weight between the predicted and measured values. It can be understood that when the covariance R(k+1) of the measurement noise is large, the gain K(k+1) decreases, indicating greater confidence in the predicted value. When the covariance Q(k+1) of the model noise is large, the gain K(k+1) increases, indicating greater confidence in the measured value.

[0116] S5. Using the information from time k+1, estimate the state at time k+1. and state covariance:

[0117]

[0118] P(k+1│k+1)=(I―K(k+1)H(k+1))P(k+1│k) (18)

[0119] in, The predicted state is based on the measured value y(k+1). The correction is as follows: P(k+1│k+1) is the updated state covariance matrix, representing the uncertainty of the updated state. In this step, the predicted value is corrected by Kalman gain to obtain a more accurate estimate of the state variables, including the estimate of the charged state. Then, the estimate of the charged state can be determined as the prediction result of the charged state, and the prediction result is output.

[0120] As can be seen, in this embodiment, by obtaining the fractional-order state-space model corresponding to the battery equivalent circuit, the state variables of the fractional-order state-space model include the state of charge; discretizing the fractional-order state-space model to obtain a discrete state-space model; adjusting the noise of the discrete state-space model to make the noise of the discrete state-space model satisfy the Gaussian distribution condition, obtaining a first discrete state-space model; and using the extended Kalman filter algorithm to predict the state of charge of the first discrete state-space model to obtain the predicted state of charge result. Since using the Kalman filter algorithm for state of charge prediction usually requires assuming that the noise of the discrete state-space model satisfies the Gaussian distribution condition, but the actual model error is difficult to satisfy the above Gaussian distribution condition, this invention, by adjusting the noise of the discrete state-space model to make the noise of the discrete state-space model satisfy the Gaussian distribution condition, can provide suitable conditions for the use of the Kalman filter algorithm. Then, by using the extended Kalman filter algorithm to predict the state of charge of the first discrete state-space model, the accuracy of the state of charge prediction can be improved. In addition, this embodiment uses a fractional-order model to model the battery. Compared with the integer-order equivalent circuit model, the fractional-order model has greater degrees of freedom, thus reducing model error and improving the accuracy of state of charge estimation.

[0121] In some implementations, the noise of the first discrete state-space model includes process noise and observation noise. In step 140, a specific implementation of predicting the state of charge of the first discrete state-space model based on the extended Kalman filter algorithm may include:

[0122] 141. Obtain the covariance matrix corresponding to the process noise and the actual parameters of the battery equivalent circuit. The actual parameters characterize the current state of the battery equivalent circuit.

[0123] In some implementations, the actual parameters may include the current input to the point battery equivalent circuit, the open-circuit voltage of the battery equivalent circuit, the state of charge, and innovation values, etc., wherein, in the fractional extended Kalman filter (EKF) algorithm, the innovation value characterizes the difference between the predicted value and the actual measured value.

[0124] 142. The elements in the covariance matrix corresponding to the state of charge are determined as target elements. The target elements represent the reciprocal of the prediction confidence of the state of charge, and the confidence represents the degree of confidence in the prediction of the state of charge.

[0125] It is understandable that, since this embodiment predicts the state of charge, and the state of charge is a third-order state variable in the fractional-order state-space model, the element corresponding to the third row and third column of the covariance matrix Q corresponding to the process noise is Q_3. 3,3 Q 3,3Specifically, it can be understood as the reciprocal of the reliability of the third-order dynamic feature of the model (i.e., the dynamics describing the SOC ampere-hour integral) in the Kalman filter algorithm.

[0126] 143. Based on the actual parameters, adjust the target elements in the first discrete state space model to obtain the second discrete state space model.

[0127] In some implementations, the actual parameter may include the current. In step 143, adjusting the target element in the first discrete state-space model according to the actual parameter to obtain the second discrete state-space model may include:

[0128] If the current magnitude does not exceed the current threshold, then obtain the current amplitude of the current.

[0129] The element value of the target element is determined based on the current amplitude, and the element value of the target element is positively correlated with the current amplitude.

[0130] If the current magnitude exceeds the current threshold, the value of the target element remains unchanged.

[0131] For example, if the current magnitude is m1 amperes (A) and the current threshold is m0 amperes, then if m1 ≤ m0, the target element Q can be determined by the magnitude of the current. 3,3 The larger the element value of the current amplitude, the greater the Q value. 3,3 The larger the element value, the better.

[0132] If m1 > m0, then the target element Q can be kept. 3,3 The element values ​​remain unchanged.

[0133] In this embodiment, to avoid drastic changes in the State of Charge (SOC) of the battery when the current is low or even zero, the target element Q is [valued / targeted] when the current is low. 3,3 The element values ​​are reduced to make the Kalman filter algorithm consider the ampere-hour integral result more reliable, so that its estimation result is closer to the ampere-hour integral result, thereby effectively improving the prediction accuracy of the state of charge when the current of the battery is small.

[0134] In some implementations, the actual parameters include the current state of charge. In step 143, the target elements in the first discrete state-space model are adjusted according to the actual parameters to obtain the second discrete state-space model. Specific implementations may include:

[0135] Determine the element value of the target element according to the current state of charge (SOC). When the current SOC is between the middle value and the upper limit value corresponding to the SOC, the current SOC is positively correlated with the element value of the target element; when the current SOC is between the lower limit value and the middle value corresponding to the SOC, the current SOC is negatively correlated with the element value of the target element.

[0136] Exemplarily, for example, if the value of the current SOC is n, the upper limit value of the SOC is 1, the lower limit value is 0, and the middle value is 0.5. If 0 < n ≤ 0.5, then the element value of the target element Q 3,3 decreases as n increases. If 0.5 < n ≤ 1, then the element value of the target element Q 3,3 increases as n increases.

[0137] In this embodiment, considering that in the extremely low and extremely high sections of the SOC, the open circuit voltage (OCV) of the battery changes relatively violently with the SOC, the accuracy of estimating the SOC using the OCV estimated by the model is relatively high. Therefore, the element value of the target element Q 3,3 can be increased, and the credibility of the ampere-hour integration can be reduced.

[0138] In some embodiments, the live parameter includes the relationship curve between the open circuit voltage of the battery corresponding to the current state of charge and the state of charge (OCV-SOC). In step 143, according to the live parameter, adjusting the target element in the first discrete state space model to obtain the second discrete state space model may specifically include:

[0139] Obtain the derivative of the relationship curve.

[0140] Determine the element value of the target element according to the derivative, and the derivative is positively correlated with the element value of the target element.

[0141] In this embodiment, considering that in the extremely low and extremely high sections of the SOC, the open circuit voltage (OCV) of the battery changes relatively violently with the SOC. Therefore, the accuracy of estimating the SOC using the OCV estimated by the model is relatively high. Therefore, determining the element value of the target element according to the derivative of the relationship curve between the open circuit voltage of the battery and the state of charge (OCV-SOC), and the derivative is positively correlated with the element value of the target element can improve the prediction accuracy of the state of charge.

[0142] In some embodiments, the live parameter includes the current innovation value. According to the live parameter, in step 143, adjusting the target element in the first discrete state space model to obtain the second discrete state space model may specifically include:

[0143] Determine the element value of the target element according to the current innovation value, and the innovation value is positively correlated with the element value of the target element.

[0144] Considering that when a sudden large SOC error occurs, the innovation value in the EKF (such as the difference between the voltage reconstructed using information from the previous moment and the measured voltage) will change significantly, the element value of the target element can be determined based on the current innovation value. The innovation value is positively correlated with the element value of the target element, which can improve the prediction accuracy of the state of charge.

[0145] 144. Predict the state of charge of the second discrete state-space model based on the extended Kalman filter algorithm.

[0146] The specific implementation of step 144 can refer to the specific implementation of step 140, which uses the extended Kalman filter algorithm to predict the state of charge of the first discrete state space model, and therefore will not be repeated here.

[0147] In some implementations, the battery equivalent circuit is used as... Figure 2 Taking the second-order R-CPE equivalent circuit shown as an example, in order to improve the prediction accuracy of the state of charge, the parameters in the EKF algorithm in this embodiment can also be dynamically adjusted according to the covariance between the state variables. For example, since there is a certain degree of positive correlation between the estimation error of the voltage drop V2 (i.e., state variable x2) of the second R-CPE branch in the battery equivalent circuit and the estimation error of the state of charge (i.e., state variable x3), the covariance terms corresponding to x2 and x3 in the EKF algorithm can be adjusted from 0 to a specified value.

[0148] In some implementations, such as Figure 3 As shown, the state of charge prediction method may further include:

[0149] 310. Obtain the current state of charge when the battery equivalent circuit is in a charging state.

[0150] 320. Determine whether the current charge state is greater than the charge state threshold.

[0151] 330. If the current state of charge is less than or equal to the state of charge threshold, the state of charge of the battery equivalent circuit is predicted based on the ampere-hour integral algorithm to obtain the predicted state of charge result.

[0152] 340. If the current state of charge is greater than the state of charge threshold, then perform the step of predicting the state of charge of the first discrete state space model based on the extended Kalman filter algorithm to obtain the predicted state of charge.

[0153] For example, the battery equivalent circuit is set to a charging state, and the current state of charge (SOC) is obtained as SOC1, with a SOC threshold of SOC0, where SOC0 is less than 0.5, specifically 0.3. If SOC1 ≤ SOC0, the SOC of the battery equivalent circuit can be predicted based on the ampere-hour integral algorithm to obtain the predicted SOC result. If SOC1 > SOC0, the process described in step 140 above, which uses the extended Kalman filter algorithm to predict the SOC of the first discrete state-space model, can be executed to obtain the predicted SOC result. Specifically, taking a SOC threshold of 0.3 as an example, when SOC < 0.3 and charging, the output SOC is estimated using a pure ampere-hour integral algorithm. However, the SOC prediction method based on the EKF algorithm in this embodiment can still run, and its output predicted SOC is forced to be synchronized with the SOC estimated by the ampere-hour integral.

[0154] In this embodiment, it is considered that, except for the case where the SOC is already in the lower range at power-on, the SOC will definitely undergo a discharge process before entering the lower range. During this process, the SOC has been estimated more accurately. Therefore, in the subsequent charging process, the ampere-hour integration method can be directly used to improve the accuracy of short-term SOC estimation, thereby improving the accuracy of state of charge prediction.

[0155] As an example, such as Figure 4 As shown, Figure 4 The diagram shows a comparison curve of SOC between the EKF-estimated SOC, the actual SOC, and the ampere-hour integral SOC estimated using the state-of-charge prediction method of this embodiment, under conditions of significant battery charge and discharge. Figure 4 In the coordinate system shown, the horizontal axis represents time, and the vertical axis represents SOC. The true SOC is the SOC obtained by using the ampere-hour integration method after compensating for the coulombic efficiency of the battery.

[0156] As an example, as shown in the figure, Figure 5 The diagram shows a comparison curve between the EKF estimation error (i.e., the error between the predicted value and the true value) of the EKF-estimated SOC obtained by the state of charge prediction method of this embodiment and the ampere-hour integral error of the SOC estimated by the ampere-hour integral method when the battery is under large charge and discharge conditions.

[0157] As an example, such as Figure 6 As shown, Figure 6 The diagram shows a comparison curve of SOC between the EKF estimated SOC, the actual SOC, and the ampere-hour integral SOC estimated using the state-of-charge prediction method of this embodiment, when the battery is under pulse charge-discharge conditions.

[0158] As an example, such as Figure 7 As shown, Figure 7 The diagram shows a comparison curve between the EKF estimation error of the EKF-estimated SOC obtained by the state of charge prediction method of this embodiment and the ampere-hour integral error of the SOC estimated using the ampere-hour integral method when the battery is under pulse charge-discharge conditions.

[0159] according to Figure 4 , Figure 5 , Figure 6 as well as Figure 7 It can be seen that when the battery is under pulse charge and discharge conditions and under large charge and discharge conditions, in the initial stage of operation of the state of charge prediction method in this embodiment, the estimated SOC gradually converges to the vicinity of the true value, and in the subsequent operation, the error between the estimated SOC value and the true value does not exceed 2%, which shows good SOC prediction performance.

[0160] Figure 8 The diagram shown is a block diagram of a state of charge prediction device according to an embodiment of the present invention. Figure 8 As shown, the state of charge prediction device 800 includes:

[0161] The model acquisition module 810 is used to acquire the fractional-order state-space model corresponding to the battery equivalent circuit. The state variables of the fractional-order state-space model include the state of charge.

[0162] Discrete processing module 820 is used to discretize the fractional-order state-space model to obtain a discrete state-space model;

[0163] The noise adjustment module 830 is used to adjust the noise of the discrete state space model so that the noise of the discrete state space model satisfies the Gaussian distribution condition, thereby obtaining the first discrete state space model.

[0164] The prediction module 840 is used to predict the state of charge of the first discrete state-space model based on the extended Kalman filter algorithm, and obtain the prediction result of the state of charge.

[0165] In another embodiment of the present invention, the noise adjustment module 830 is specifically used for:

[0166] Add noise estimates to the state variables of the discrete state-space model;

[0167] The noise estimate is updated based on the random walk algorithm, and the noise of the discrete state-space model is generated based on the updated noise estimate.

[0168] In another embodiment of the present invention, the noise of the first discrete state-space model includes process noise, and the prediction module 840 is specifically used for:

[0169] Obtain the covariance matrix corresponding to the process noise and the actual parameters of the battery equivalent circuit. The actual parameters characterize the current state of the battery equivalent circuit.

[0170] The elements in the covariance matrix corresponding to the state of charge are determined as target elements, which represent the reciprocal of the prediction confidence of the state of charge.

[0171] Based on the actual parameters, the target elements in the first discrete state space model are adjusted to obtain the second discrete state space model.

[0172] The extended Kalman filter algorithm is used to predict the state of charge of the second discrete state-space model.

[0173] In another embodiment of the present invention, the real-time parameters include the current current, and the prediction module 840 is further configured to:

[0174] If the current magnitude does not exceed the current threshold, then obtain the current amplitude of the current.

[0175] The element value of the target element is determined based on the current amplitude, and the element value of the target element is positively correlated with the current amplitude.

[0176] If the current magnitude exceeds the current threshold, the value of the target element remains unchanged.

[0177] In another embodiment of the present invention, the real-time parameters include the current state of charge, and the prediction module 840 is further configured to:

[0178] Based on the current state of charge, determine the element value of the target element. When the current state of charge is between the middle and upper limits of the corresponding state of charge, the current state of charge is positively correlated with the element value of the target element; when the current state of charge is between the lower limit and the middle limit of the corresponding state of charge, the current state of charge is negatively correlated with the element value of the target element.

[0179] In another embodiment of the present invention, the real-time parameters include the relationship curve between the battery open-circuit voltage corresponding to the current state of charge and the state of charge. The prediction module 840 is further used for:

[0180] Obtain the derivative corresponding to the relationship curve;

[0181] The element value of the target element is determined by the derivative, and the derivative is positively correlated with the element value of the target element.

[0182] In another embodiment of the present invention, the real-time parameters include the current information value, and based on the real-time parameters, the prediction module 840 is further configured to:

[0183] Based on the current information value, determine the element value of the target element. The information value is positively correlated with the element value of the target element.

[0184] In another embodiment of the present invention, the prediction module 840 is further configured to:

[0185] When the battery equivalent circuit is in a charging state, obtain the current state of charge.

[0186] If the current state of charge is less than or equal to the state of charge threshold, the state of charge of the battery equivalent circuit is predicted based on the ampere-hour integral algorithm to obtain the predicted state of charge result.

[0187] If the current state of charge is greater than the state of charge threshold, then the step of performing state of charge prediction on the first discrete state space model based on the extended Kalman filter algorithm is executed to obtain the predicted state of charge.

[0188] In another embodiment of the present invention, the battery equivalent circuit is a second-order R-CPE equivalent circuit containing fractional-order constant-phase elements. The fractional-order state-space model includes first-order state variables, second-order state variables, and third-order state variables. The first-order state variables characterize the voltage drop corresponding to the electrochemical polarization resistance in the second-order R-CPE equivalent circuit, the second-order state variables characterize the voltage drop corresponding to the concentration difference polarization resistance in the second-order R-CPE equivalent circuit, and the third-order state variables characterize the state of charge.

[0189] In another embodiment of the present invention, the discrete state-space model includes fourth-order state variables, which characterize the noise estimates of the discrete state-space model.

[0190] The specific implementation process of the functions and roles of each module in the above device can be found in the implementation process of the corresponding steps of the state of charge prediction method in the above embodiments, and will not be repeated here.

[0191] Figure 9 The diagram shown is a block diagram of an electronic device 900 provided in an embodiment of the present invention.

[0192] Reference Figure 9 The electronic device 900 includes a processing component 910, which further includes one or more processors, and memory resources represented by a memory 920 for storing instructions, such as application programs, that can be executed by the processing component 910. The application programs stored in the memory 920 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 910 is configured to execute instructions to perform the aforementioned state of charge prediction method.

[0193] Electronic device 900 may also include a power supply component configured to perform power management of electronic device 900, a wired or wireless network interface configured to connect electronic device 900 to a network, and an input / output (I / O) interface. Electronic device 900 can operate on an operating system stored in memory 920, such as Windows Server™, Mac OSX™, Unix™, Linux™, FreeBSD™, or similar.

[0194] A non-transitory computer-readable storage medium, when the instructions in the storage medium are executed by the processor of the electronic device 900, enables the electronic device 900 to perform the state of charge prediction method.

[0195] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0196] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0197] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0198] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0199] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0200] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program verification codes.

[0201] Furthermore, it should be noted that the combination of the various technical features in this case is not limited to the combination methods described in the claims of this case or the combination methods described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless they contradict each other.

[0202] It should be noted that the above examples are merely specific embodiments of the present invention, and the present invention is obviously not limited to the above embodiments, with many similar variations. All modifications that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should fall within the protection scope of this invention.

[0203] It should be understood that the terms "first," "second," etc., mentioned in the embodiments of the present invention are merely for the purpose of more clearly describing the use of the technical solutions in the embodiments of the present invention, and are not intended to limit the scope of protection of the present invention.

[0204] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the state of charge, characterized in that, include: Obtain the fractional-order state-space model corresponding to the battery equivalent circuit, wherein the state variables of the fractional-order state-space model include the state of charge. The fractional-order state-space model is discretized to obtain a discrete state-space model; The noise of the discrete state space model is adjusted so that the noise of the discrete state space model satisfies the Gaussian distribution condition, thus obtaining the first discrete state space model. The state of charge is predicted based on the extended Kalman filter algorithm of the first discrete state-space model, and the prediction result of the state of charge is obtained.

2. The method according to claim 1, characterized in that, The noise adjustment process for the discrete state-space model includes: Noise estimates are added to the state variables of the discrete state-space model. The noise estimate is updated based on the random walk algorithm, and the noise of the discrete state space model is generated based on the updated noise estimate.

3. The method according to claim 1, characterized in that, The noise in the first discrete state-space model includes process noise. The step of predicting the state of charge of the first discrete state-space model based on the extended Kalman filter algorithm includes: Obtain the covariance matrix corresponding to the process noise and the actual parameters of the battery equivalent circuit, wherein the actual parameters characterize the current state of the battery equivalent circuit; The element in the covariance matrix corresponding to the state of charge is determined as the target element, and the target element represents the reciprocal of the prediction confidence of the state of charge. Based on the actual parameters, the target elements in the first discrete state space model are adjusted to obtain the second discrete state space model. The state of charge of the second discrete state-space model is predicted based on the extended Kalman filter algorithm.

4. The method according to claim 3, characterized in that, The real-world parameters include the current current. The step of adjusting the target elements in the first discrete state-space model based on the real-world parameters to obtain the second discrete state-space model includes: If the magnitude of the current does not exceed the current threshold, then the current amplitude of the current is obtained; The element value of the target element is determined based on the current amplitude, and the element value of the target element is positively correlated with the current amplitude. If the magnitude of the current exceeds the current threshold, the element value of the target element remains unchanged.

5. The method according to claim 3, characterized in that, The real-world parameters include the current state of charge. Adjusting the target elements in the first discrete state-space model based on the real-world parameters to obtain the second discrete state-space model includes: Based on the current state of charge, the element value of the target element is determined, wherein when the current state of charge is between the intermediate value and the upper limit value corresponding to the state of charge, the current state of charge is positively correlated with the element value of the target element; when the current state of charge is between the lower limit value and the intermediate value corresponding to the state of charge, the current state of charge is negatively correlated with the element value of the target element.

6. The method according to claim 3, characterized in that, The real-world parameters include the relationship curve between the battery open-circuit voltage and the current state of charge. The step of adjusting the target elements in the first discrete state-space model based on the real-world parameters to obtain a second discrete state-space model includes: Obtain the derivative corresponding to the relationship curve; The element value of the target element is determined based on the derivative, and the derivative is positively correlated with the element value of the target element.

7. The method according to claim 3, characterized in that, The real-time parameters include the current information value. Adjusting the target elements in the first discrete state-space model based on the real-time parameters to obtain the second discrete state-space model includes: Based on the current information value, the element value of the target element is determined, and the information value is positively correlated with the element value of the target element.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: When the battery equivalent circuit is in a charging state, the current state of charge is obtained; If the current state of charge is less than or equal to the state of charge threshold, then the state of charge of the battery equivalent circuit is predicted based on the ampere-hour integral algorithm to obtain the predicted state of charge. If the current state of charge is greater than the state of charge threshold, then the step of performing state of charge prediction on the first discrete state-space model based on the extended Kalman filter algorithm is executed to obtain the prediction result of the state of charge.

9. The method according to any one of claims 1 to 7, characterized in that, The battery equivalent circuit is a second-order R-CPE equivalent circuit containing fractional-order constant-phase elements. The fractional-order state-space model includes first-order state variables, second-order state variables, and third-order state variables. The first-order state variables represent the voltage drop corresponding to the electrochemical polarization resistance in the second-order R-CPE equivalent circuit, the second-order state variables represent the voltage drop corresponding to the concentration difference polarization resistance in the second-order R-CPE equivalent circuit, and the third-order state variables represent the state of charge.

10. The method according to claim 9, characterized in that, The discrete state-space model includes fourth-order state variables, which represent the noise estimates of the discrete state-space model.

11. A state of charge prediction device, characterized in that, include: The model acquisition module is used to acquire the fractional-order state-space model corresponding to the battery equivalent circuit, wherein the state variables of the fractional-order state-space model include the state of charge. The discrete processing module is used to discretize the fractional-order state-space model to obtain a discrete state-space model. The noise adjustment module is used to adjust the noise of the discrete state space model so that the noise of the discrete state space model satisfies the Gaussian distribution condition, thereby obtaining the first discrete state space model. The prediction module is used to predict the state of charge of the first discrete state-space model based on the extended Kalman filter algorithm, and obtain the prediction result of the state of charge.

12. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that, When the executable instructions are executed by the processor, they implement the state of charge estimation method as described in any one of claims 1 to 10.

13. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the state of charge estimation method according to any one of claims 1 to 10.

Images