A battery parameter online identification method and system

Abstract

The present invention discloses a battery parameter online identification method and system, belonging to the field of battery technology. A forgetting matrix and an information matrix that is an inverse matrix of a covariance matrix are introduced, and the information matrix is ​​split into an unstimulated information subspace matrix and an stimulated information subspace matrix. Only the stimulated information subspace matrix is ​​forgotten. Based on the covariance matrix update formula in the existing recursive least squares algorithm with a forgetting factor, the forgetting matrix is ​​multiplied by the information matrix to replace the covariance matrix in the original covariance matrix update formula. This is used as a directional forgetting mechanism, and the error covariance matrix update formula #imgabs0# adopted by the present invention is derived. The method can selectively forget old information components that can be replaced by the new information input at the current moment, while other old information components are not affected. This avoids the occurrence of covariance explosion, greatly improves robustness, and makes the identification accuracy of battery parameters higher.

Description

Technical Field

[0001] The present invention belongs to the field of battery technology, and more specifically, relates to a battery parameter online identification method and system. Background Art

[0002] In the green, low-carbon, and sustainable development of human society, battery technology, as a key energy storage solution, plays a vital role. Whether it is an energy storage battery or a power battery, real-time and accurate battery state estimation is an important prerequisite for safe, stable, and efficient battery operation. However, core battery states, including state of charge, state of health, and power state, cannot be directly measured and can only be estimated through related physical quantities. Among them, state estimation methods based on voltage, current, and temperature are the most common. Battery parameter identification is an essential step in the state estimation method based on voltage, current, and temperature. The accuracy of battery parameter identification directly affects the accuracy of battery state estimation. Therefore, it is of great significance to study a battery parameter identification method.

[0003] A commonly used online battery parameter identification method is the recursive least squares algorithm with a forgetting factor. The algorithm has the advantages of good convergence and low real-time computational overhead, making it highly suitable for the application requirements of online battery model parameter identification. This method typically sets independent time-varying forgetting factors for different model parameters, adaptively adjusting them based on real-time system changes to increase the weight of new data, reduce the influence of old data, and quickly track the latest state of the target system. However, when the current excitation is non-continuous, a covariance explosion occurs, resulting in periodic fluctuations in the results and poor robustness, leading to low battery parameter identification accuracy. Summary of the Invention

[0004] In response to the above defects or improvement needs of the prior art, the present invention provides a method and system for online identification of battery parameters to solve the technical problem of low accuracy in battery parameter identification in the prior art.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for online identification of battery parameters, comprising:

[0006] S1. At the current sampling time k, perform a target dynamic operating condition test on the battery to obtain the battery current I under the target dynamic operating condition. k and terminal voltage U k , and then get a k and b k ; a k and b kare respectively used as the input vector and output value of the battery equivalent circuit model at time k; wherein the battery equivalent circuit model is a model obtained by discretizing the r-order equivalent circuit model of the battery; r is a positive integer; a k For-(U k-1 -U oc,k-1 ),-(U k-2 -U oc,k-2 ),…,-(U k-r -U oc,k-r ) and I k ,I k-1 ,…,I k-r The 2r+1 dimensional vector formed; b k =U k -U oc,k ;U oc,k is the open circuit voltage of the battery under the target dynamic condition at time k; k ≥ r;

[0007] S2. Calculate the measurement error of the battery equivalent circuit model at time k in, is the battery parameter estimation vector at time k-1, specifically a 2r+1 dimensional vector, which is constructed based on the estimated values ​​of each battery parameter to be identified in the battery equivalent circuit model at time k-1;

[0008] Using the error covariance matrix update formula Calculate the error covariance matrix P at time k k , and then calculate the system gain vector K at time k k =P k a k ;in, is the error covariance matrix after directed forgetting at time k-1, specifically: λ is the forgetting factor;

[0009] S3. Calculate the battery parameter estimation vector at time k Then, the estimated values ​​of the battery parameters to be identified at time k are obtained.

[0010] Further preferably, the discretization process includes:

[0011] Discretize the system transfer function of the r-order equivalent circuit model to obtain a discretized system transfer function;

[0012] The discretized system transfer function is converted into a discrete linear expression representing the battery voltage-current relationship as the above-mentioned battery equivalent circuit model.

[0013] Further preferably, when r=1, the battery parameters to be identified include: the battery's ohmic internal resistance, polarization resistance, and polarization capacitance.

[0014] Further preferably, the above-mentioned r-order equivalent circuit model is a first-order Thevenin equivalent circuit model, and its expression is:

[0015]

[0016] Where C1 is the polarization capacitance of the battery; U1 is the polarization voltage of the battery; t represents time; R1 is the polarization resistance of the battery; I is the current of the battery; U oc is the open circuit voltage of the battery; R0 is the ohmic internal resistance of the battery; U is the terminal voltage of the battery.

[0017] Further preferably, the battery equivalent circuit model is:

[0018]

[0019] Among them, y k is the output value of the battery equivalent circuit model at time k, specifically U k -U oc,k ;φ k =[-y k-1 x k x k-1 ] is the input vector of the battery equivalent circuit model at time k; x k For I k θ k =[m 1,k n 1,k n 2,k ] is the battery parameter estimation vector at time k; e k is the measurement error of the battery equivalent circuit model at time k; m 1,k 、n 1,k and n 2,k The following relations are satisfied: n 1,k =R 0,k , T is the sampling period; R 0,k is the ohmic internal resistance of the battery to be identified at time k; R 1,k is the polarization resistance of the battery to be identified at time k; C 1,k is the polarization capacitance of the battery to be identified at time k.

[0020] Further preferably, the open circuit voltage U of the battery at time k under the target dynamic operating condition is oc,k Obtained through the following methods:

[0021] Get the battery's state of charge (SoC) at time k under the target dynamic conditions k, and according to the mapping relationship between the battery open circuit voltage and the state of charge, the open circuit voltage U of the battery at the target dynamic condition at time k is obtained oc,k ;

[0022] The mapping relationship between the battery open circuit voltage and the state of charge is obtained by performing an open circuit voltage test on the battery under target dynamic operating conditions.

[0023] In a second aspect, the present invention provides a battery parameter online identification system, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor executes the battery parameter online identification method provided by the first aspect of the present invention when executing the computer program.

[0024] In a third aspect, the present invention provides a battery management system, comprising: the battery parameter online identification system provided in the second aspect of the present invention.

[0025] In a fourth aspect, the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the battery parameter online identification method provided in the first aspect of the present invention.

[0026] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0027] 1. The present invention provides a method for online identification of battery parameters. Considering that the old information in the existing recursive least squares algorithm with a forgetting factor is forgotten indiscriminately, the identification results obtained are prone to periodic fluctuations when there is non-continuous excitation. The present invention introduces a directional forgetting mechanism to optimize the existing forgetting mechanism, and designs a directional forgetting recursive least squares algorithm to perform online identification of battery parameters. The forgetting matrix and the information matrix that is the inverse matrix of the covariance matrix are introduced, and the information matrix is ​​split into an unstimulated information subspace matrix and an stimulated information subspace matrix, and only the stimulated information subspace matrix is ​​forgotten. Based on the covariance matrix update formula in the existing recursive least squares algorithm with a forgetting factor, the covariance matrix in the original covariance matrix update formula is replaced by multiplying the forgetting matrix by the information matrix, and this is used as a directional forgetting mechanism to derive the error covariance matrix update formula adopted by the present invention. Through this design, the forgetting factor is changed from acting on the entire covariance matrix to acting selectively on the excited information subspace. Then, according to the new information input at the current moment, the old information components that can be replaced by it can be selectively forgotten, while other old information components are not affected. This avoids the occurrence of covariance explosion, greatly improves the robustness, and makes the identification accuracy of battery parameters higher.

[0028] 2. Furthermore, the battery parameter online identification method provided by the present invention preferably adopts a first-order Thevenin equivalent circuit model, which has a simple structure and can simulate the dynamic and static voltage characteristics of the battery, thereby facilitating online parameter identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic structural diagram of an r-order equivalent circuit model of a battery provided in an embodiment of the present invention;

[0030] Figure 2 A flowchart of a method for online identification of battery parameters provided by an embodiment of the present invention;

[0031] Figure 3 A schematic diagram of the structure of a first-order Thevenin model of a battery provided by an embodiment of the present invention;

[0032] Figure 4 A schematic diagram of a current-voltage curve under dynamic working conditions provided by an embodiment of the present invention;

[0033] Figure 5 The OCV-SoC curve diagram of the battery provided by the embodiment of the present invention;

[0034] Figure 6 Comparison of online identification result curves of ohmic internal resistance, polarization resistance, and time constant of a liquid metal battery provided in an embodiment of the present invention using the RLSDF algorithm and the RLSFF algorithm under DST conditions. DETAILED DESCRIPTION

[0035] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0036] To achieve the above objectives, in a first aspect, the present invention provides a method for online identification of battery parameters, comprising:

[0037] S1. At the current sampling time k, perform a target dynamic operating condition test on the battery to obtain the battery current I under the target dynamic operating condition. k and terminal voltage U k , and then get a k and b k ; a k and b k are respectively used as the input vector and output value of the battery equivalent circuit model at time k; wherein the battery equivalent circuit model is a model obtained by discretizing the r-order equivalent circuit model of the battery; r is a positive integer; ak For-(U k-1 -U oc,k-1 ),-(U k-2 -U oc,k-2 ),…,-(U k-r -U oc,k-r ) and I k ,I k-1 ,…,I k-r The 2r+1 dimensional vector formed; b k =U k -U oc,k ;U oc,k is the open circuit voltage of the battery under the target dynamic condition at time k; k ≥ r;

[0038] The discretization process includes: discretizing the system transfer function of the r-order equivalent circuit model to obtain a discretized system transfer function; and converting the discretized system transfer function into a discrete linear expression representing the battery voltage-current relationship as the battery equivalent circuit model.

[0039] Specifically, without considering the actual measurement error, the battery equivalent circuit model is:

[0040]

[0041] Among them, y k As the output value of the battery equivalent circuit model at time k, specifically y k =U k -U oc,k ;-y k-1 ,-y k-2 ,…,-y k-n and x k ,x k-1 ,…,x k-n The 2r+1 dimensional vector formed is used as the input vector θ of the battery equivalent circuit model at time k k ;-m 1,k ,-m 2,k ,…,-m r,k and -n 1,k ,-n 2,k ,…,-n r+1,k The 2n+1 dimensional vector φ k ;y k 、φ k and θ k The corresponding relationship satisfies It should be noted that m i,k and n j,k It is constructed based on the estimated values ​​of each battery parameter to be identified in the battery equivalent circuit model, and its specific form is determined by the specific r-order equivalent circuit model and the sampling period T.

[0042] When taking into account the actual measurement error, the battery equivalent circuit model is: Among them, e k is the measurement error of the battery equivalent circuit model at time k.

[0043] In the present invention, y k =b k ,θ k =a k , in, is the battery parameter estimation vector at time k to be identified.

[0044] It should be noted that the above-mentioned target dynamic operating conditions may be dynamic stress conditions (DST), highway fuel economy test (HWFET), China light vehicle test conditions (CLTC), New York city conditions (NYCC), urban dynamometer driving cycle conditions (UDDS), etc., and are not limited here.

[0045] It should be noted that in an optional embodiment, the open circuit voltage U of the battery under the target dynamic operating condition at time k is oc,k It can be obtained in the following ways:

[0046] Get the battery's state of charge (SoC) at time k under the target dynamic conditions k , and according to the mapping relationship between the battery open circuit voltage and the state of charge, the open circuit voltage U of the battery at the target dynamic condition at time k is obtained oc,k ;

[0047] The mapping relationship between the battery open circuit voltage and the state of charge is obtained by testing the battery open circuit voltage under the target dynamic working conditions. k It can be calculated by the ampere-hour integration method.

[0048] S2. Calculate the measurement error of the battery equivalent circuit model at time k in, is a battery parameter estimation vector at time k-1, specifically a 2r+1-dimensional vector, constructed based on the estimated values ​​of each battery parameter to be identified in the battery equivalent circuit model at time k-1;

[0049] Using the error covariance matrix update formula Calculate the error covariance matrix P at time k k , and then calculate the system gain vector K at time k k =P k a k ;in, is the error covariance matrix after directed forgetting at time k-1, specifically: λ is the forgetting factor;

[0050] S3. Calculate the battery parameter estimation vector at time k Then, the estimated values ​​of the battery parameters to be identified at time k are obtained.

[0051] It should be noted that the r-order equivalent circuit model is usually the r-order RC equivalent circuit model of the battery, as shown in the following example: Figure 1 As shown, the model consists of an open circuit voltage source U oc , an ohmic internal resistance R0 and r polarization resistors R i and polarization capacitance C i It should be noted that r is a positive integer, and is generally set to 1, 2, 3, etc. according to user needs.

[0052] In an optional embodiment, when r=1, the battery parameters to be identified include: the battery's ohmic internal resistance, polarization resistance, and polarization capacitance. The r-order equivalent circuit model is a first-order Thevenin equivalent circuit model, and its expression is:

[0053]

[0054] Where C1 is the polarization capacitance of the battery; U1 is the polarization voltage of the battery; t represents time; R1 is the polarization resistance of the battery; I is the current of the battery; U oc is the open circuit voltage of the battery; R0 is the ohmic internal resistance of the battery; U is the terminal voltage of the battery.

[0055] At this time, the above battery equivalent circuit model is:

[0056]

[0057] Among them, y k is the output value of the battery equivalent circuit model at time k, specifically U k -U oc,k ;φ k =[-y k-1 x k x k-1 ] is the input vector of the battery equivalent circuit model at time k; x k For I k θ k =[m 1,k n 1,k n 2,k ] is the battery parameter estimation vector at time k; e k is the measurement error of the battery equivalent circuit model at time k; m 1,k 、n 1,kand n 2,k The following relations are satisfied: n 1,k =R 0,k , T is the sampling period; R 0,k is the ohmic internal resistance of the battery to be identified at time k; R 1,k is the polarization resistance of the battery to be identified at time k; C 1,k is the polarization capacitance of the battery to be identified at time k.

[0058] It should be noted that the above is an example of the battery equivalent circuit model expression when the first-order Thevenin equivalent circuit model is used. The battery equivalent circuit model expression examples when other models are used are similar to it, and the only difference is that φ k and θ k The specific composition of φ k and θ k It can be derived from the corresponding r-order equivalent circuit model adopted through the above discretization process. The acquisition method is the same as the acquisition process when the first-order Thevenin equivalent circuit model is adopted, and will not be repeated here.

[0059] In summary, the present invention provides a battery parameter identification method that introduces a directed forgetting recursive mechanism. During the model parameter identification process, the forgetting factor no longer acts indiscriminately on the entire information matrix. Instead, it affects the update of the information matrix in the form of a forgetting matrix, thereby increasing the information content of the information matrix. Based on the new information input at the current moment, the old information components that can be replaced by it are selectively forgotten, while other old information components are unaffected. This reduces the identification method's sensitivity to noise, improves the robustness of the identification, and thus improves the identification accuracy.

[0060] In order to further illustrate the battery parameter online identification method provided by the present invention, a specific embodiment is described in detail below:

[0061] This example uses a Li-Bi liquid metal battery as the research object. The battery has an actual capacity of 52Ah, an operating temperature of 550°C, and charge and discharge cut-off voltages of 1.2V and 0.5V, respectively. Based on test data from dynamic operating conditions, the method proposed in this invention is used to perform online parameter identification of the battery model. The battery parameters to be identified in this example include the battery's ohmic internal resistance, polarization resistance, and polarization capacitance, using a first-order Thevenin model.

[0062] like Figure 2 As shown in the figure, the entire online identification process of battery parameters is as follows:

[0063] 1) Before formally executing the online identification process, construct a battery equivalent circuit model;

[0064] Establish a first-order Thevenin model of liquid metal battery, such as Figure 3 As shown: The model consists of an open circuit voltage source U oc , an ohmic internal resistance R0, and an RC parallel network consisting of a polarized resistor R1 and a polarized capacitor C1. The mathematical expression of the model is as follows:

[0065]

[0066] Where, t represents time; I is the current of the battery; U oc is the open circuit voltage of the battery; U is the terminal voltage of the battery.

[0067] Establish the system transfer function based on this model. If the current is the input and the voltage of the passive network is the output, the continuous transfer function H(s) of the system is:

[0068]

[0069] Discretize the continuous system. Use the first-order forward difference method to discretize the transfer function. Assuming the sampling period is T, the discretized transfer function is as follows:

[0070]

[0071] Where z is the discretized z-domain complex variable, T is the sampling period, and H(z) is the discretized system transfer function;

[0072] m1, n1, and n2 are all intermediate parameters for solving model parameters. Their specific values ​​are After identifying the intermediate parameters m1, n1, and n2, we can Solve for the equivalent circuit model parameters R0, R1, and C1.

[0073] The discrete form of the voltage-current relationship is further obtained as: k =-m1y k-1 +n1x k +n2x k-1 .

[0074] Rewriting it into vector form and taking into account the actual measurement error, the battery equivalent circuit model is obtained, which is:

[0075] y k =φ k θ k +e k

[0076] Among them, y k is the output value of the battery equivalent circuit model at time k, specifically U k -U oc,k ;φk =[-y k-1 x k x k-1 ] is the input vector of the battery equivalent circuit model at time k; x k For I k θ k =[m 1,k n 1,k n 2,k ] is the battery parameter estimation vector at time k; e k is the measurement error of the battery equivalent circuit model at time k; m 1,k 、n 1,k and n 2,k The following relations are satisfied: n 1,k =R 0,k , T is the sampling period; R 0,k is the ohmic internal resistance of the battery to be identified at time k; R 1,k is the polarization resistance of the battery to be identified at time k; C 1,k is the polarization capacitance of the battery to be identified at time k.

[0077] 2) During the formal online identification process, at the current time k, the battery is tested under the target dynamic operating condition to obtain the battery current I under the target dynamic operating condition. k and terminal voltage U k , and then get a k and b k .

[0078] In this embodiment, the battery is first charged at a constant current of 0.2C to a cut-off voltage of 1.2V, and then constant voltage charging is performed until the charge rate drops to 0.05C to ensure that the battery reaches a full state. Subsequently, after a period of rest, the battery voltage gradually stabilizes. The battery is subjected to a dynamic stress condition (DST) for 2 hours using a Blue Power CT3001D battery tester, and the terminal voltage and current data of the battery are recorded at a sampling frequency of 1Hz. The current-voltage curve under the dynamic condition is shown in Figure 2. Figure 4 shown.

[0079] Prior to this, the circuit was tested for open circuit voltage to determine the mapping relationship between the battery open circuit voltage and the state of charge, which is recorded as the battery OCV-SoC relationship. The liquid metal battery was charged and discharged at a constant current rate of 0.05C, and the voltage and SoC were recorded in real time. The test results are as follows: Figure 5 As shown, the two dotted lines correspond to the voltage curves during the charge and discharge process. The final battery OCV-SoC relationship curve is obtained by averaging the voltage curves during the charge and discharge process, which is shown as a solid line in the figure. The functional relationship between OCV and SoC is processed using a table lookup method.

[0080] In this embodiment, a k =[-(U k-1 -U oc,k-1 )I k I k-1 ]; among them, U k-1 is the terminal voltage of the battery at time k-1; U oc,k-1 is the open circuit voltage of the battery at time k-1; I k is the battery current at time k; I k-1 is the battery current at time k.

[0081] b k =U k -U oc,k ;U oc,k is the open circuit voltage of the battery at time k; in this embodiment, the open circuit voltage of the battery at time k is U oc,k Obtained through the following methods:

[0082] Get the battery state of charge SoC at time k k , and according to the mapping relationship between the battery open circuit voltage and the state of charge, the open circuit voltage U of the battery at time k is obtained oc,k ;

[0083] The mapping relationship between the battery open circuit voltage and the state of charge is obtained by performing an open circuit voltage test on the battery. k It can be solved by the ampere-hour integration method, which is calculated by integrating the current over time.

[0084] 3) Use the error covariance matrix update formula to calculate the error covariance matrix at time k Then calculate the system gain vector K at time k k =P k a k ;in, is the error covariance matrix after directed forgetting at time k-1, specifically: λ is the forgetting factor;

[0085] It should be noted that the covariance matrix update formula in the original recursive least squares algorithm with forgetting factor is: Under this covariance matrix update formula, when the current excitation is non-continuous, covariance explosion will occur, resulting in periodic fluctuations in the results, poor robustness, and low accuracy in battery parameter identification.

[0086] In order to solve this problem, the present invention modifies it by changing the error covariance matrix P k The update process is changed to the information matrix R k The form is:

[0087]

[0088] P k With R k The information matrix R at time k-1 is inverse to each other. k-1 Split into unstimulated information subspace R 1,k-1 and the stimulated information subspace R 2,k-1 , by distinguishing R 1,k-1 and R 2,k-1 , realize the selective directional forgetting by forgetting factor λ, specifically:

[0089] Solve the stimulated information subspace R 2,k-1 , we get the following formula:

[0090]

[0091] Then we get the information matrix R of directed forgetting k The update formula is:

[0092]

[0093] R k The update process of the error covariance matrix obtained by inversion is:

[0094]

[0095] Among them, Λ is the forgetting matrix; is the error covariance matrix after directed forgetting at time k-1, specifically:

[0096]

[0097] In addition, this step also calculates the measurement error of the battery equivalent circuit model at time k in, is the battery parameter estimation vector at time k-1; the battery parameter estimation vector is a vector composed of the estimated values ​​of each battery parameter to be identified in the battery equivalent circuit model;

[0098] 3) Calculate the battery parameter estimation vector at time k Then, the estimated values ​​of the battery parameters to be identified at the current time k are obtained.

[0099] It should be noted that, during initialization, taking r=1 as an example, when k=1, P0, U0, U oc,0 All use preset initial values; in an optional implementation, P0, U0, Uoc,0 The corresponding initial values ​​are [0 0 0] T , 10 3 I, 0, 0.

[0100] The present invention refers to the above-mentioned solution algorithm for battery parameter identification as the directional forgetting recursive least squares algorithm (RLSDF). The RLSDF provided by the present invention is compared with the traditional recursive least squares algorithm with forgetting factor (RLSFF) for battery parameter identification, and the following results are obtained: Figure 6 The comparison results are shown. Figure 6 It can be seen that, thanks to the directed forgetting mechanism, the battery parameters identified by the online battery parameter identification method provided by the present invention are more stable compared with RLSFF. In RLSFF, the forgetting of old information is indiscriminate, and the current step will always lead to the same "mutation-convergence" process in the parameter estimation results, so the parameter estimation value will show periodic changes. In the present invention, due to the addition of the directed forgetting mechanism, the information stored in the information matrix is ​​richer, which is mathematically reflected in the information matrix having a larger rank, which can avoid excessive gain, thereby enhancing the robustness of the parameter estimation algorithm. From Figure 6 It can be seen that the battery model parameters identified by the battery parameter online identification method provided by the present invention will not show severe jitter phenomenon, which is particularly evident in the identification of the time constant. The present invention effectively limits the continuous increase of the gain under non-continuous excitation conditions by optimizing the update mechanism of the covariance matrix in the RLS algorithm, thereby achieving the purpose of enhancing the robustness of the algorithm and maintaining the stability of the parameter estimation results. According to the results of the preferred embodiment, compared with the traditional RLSFF algorithm, the battery parameter online identification method proposed by the present invention can provide more stable online parameter identification results, which is beneficial to enhancing the convergence performance of the filter or observer in battery state estimation.

[0101] In a second aspect, the present invention provides a battery parameter online identification system, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor executes the battery parameter online identification method provided by the first aspect of the present invention when executing the computer program.

[0102] The related technical solution is the same as the battery parameter online identification method provided in the first aspect of the present invention, and will not be described in detail here.

[0103] In a third aspect, the present invention provides a battery management system, comprising: the battery parameter online identification system provided in the second aspect of the present invention.

[0104] The related technical solution is the same as the battery parameter online identification system provided in the second aspect of the present invention, and will not be described in detail here.

[0105] In a fourth aspect, the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the battery parameter online identification method provided in the first aspect of the present invention.

[0106] The related technical solution is the same as the battery parameter online identification method provided in the first aspect of the present invention, and will not be described in detail here.

[0107] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A battery parameter online identification method, characterized in that: include: S1. At the current sampling time k, perform a target dynamic operating condition test on the battery to obtain the battery current I under the target dynamic operating condition. k and terminal voltage U k , and then get a k and b k ; a k and b k are respectively used as the input vector and output value of the battery equivalent circuit model at time k; wherein, the battery equivalent circuit model is a model obtained by discretizing the r-order equivalent circuit model of the battery; r is a positive integer; a k For-(U k-1 -U oc,k-1 ),-(U k-2 -U oc,k-2 ),…,-(U k-r -U oc,k-r ) and I k ,I k-1 ,…,I k-r The 2r+1 dimensional vector formed; b k =U k -U oc,k ;U oc,k is the open circuit voltage of the battery under the target dynamic condition at time k; k ≥ r; S2. Calculate the measurement error of the battery equivalent circuit model at time k is a battery parameter estimation vector at time k-1, specifically a 2r+1-dimensional vector, constructed based on the estimated values ​​of each battery parameter to be identified in the battery equivalent circuit model at time k-1; Using the error covariance matrix update formula Calculate the error covariance matrix P at time k k , and then calculate the system gain vector K at time k k =P k a k ; is the error covariance matrix after directed forgetting at time k-1, specifically: λ is the forgetting factor; S3. Calculate the battery parameter estimation vector at time k Then, the estimated values ​​of the battery parameters to be identified at time k are obtained.

2. The battery parameter online identification method according to claim 1, characterized in that: The discretization process includes: Discretizing the system transfer function of the r-order equivalent circuit model to obtain a discretized system transfer function; The discretized system transfer function is converted into a discrete linear expression representing the battery voltage-current relationship as the battery equivalent circuit model.

3. The battery parameter online identification method according to claim 2, characterized in that: When r=1, the battery parameters to be identified include: the battery's ohmic internal resistance, polarization resistance, and polarization capacitance.

4. The battery parameter online identification method according to claim 3, characterized in that: The r-order equivalent circuit model is a first-order Thevenin equivalent circuit model, and its expression is: Where C1 is the polarization capacitance of the battery; U1 is the polarization voltage of the battery; t represents time; R1 is the polarization resistance of the battery; I is the current of the battery; U oc is the open circuit voltage of the battery; R0 is the ohmic internal resistance of the battery; U is the terminal voltage of the battery.

5. The battery parameter online identification method according to claim 4, characterized in that: The battery equivalent circuit model is: Among them, y k is the output value of the battery equivalent circuit model at time k, specifically U k -U oc,k ;φ k =[-y k-1 x k x k-1 ] is the input vector of the battery equivalent circuit model at time k; x k For I k θ k =[m 1,k n 1,k n 2,k ] is the battery parameter estimation vector at time k; e k is the measurement error of the battery equivalent circuit model at time k; m 1,k 、n 1,k and n 2,k The following relations are satisfied: n 1,k =R 0,k , T is the sampling period; R 0,k is the ohmic internal resistance of the battery to be identified at time k; R 1,k is the polarization resistance of the battery to be identified at time k; C 1,k is the polarization capacitance of the battery to be identified at time k.

6. The method for online identification of battery parameters according to any one of claims 1 to 5, characterized in that: The open circuit voltage U of the battery under the target dynamic working condition at time k oc,k Obtained through the following methods: Get the battery's state of charge (SoC) at time k under the target dynamic conditions k , and according to the mapping relationship between the battery open circuit voltage and the state of charge, the open circuit voltage U of the battery at the target dynamic condition at time k is obtained oc,k ; The mapping relationship between the battery open circuit voltage and the state of charge is obtained by performing an open circuit voltage test on the battery under target dynamic operating conditions.

7. A battery parameter online identification system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method for online identification of battery parameters according to any one of claims 1 to 6 is executed.

8. A battery management system, characterized in that: include: The battery parameter online identification system according to claim 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the battery parameter online identification method according to any one of claims 1 to 6.

Images