Battery parameter identification method, device, computer equipment and storage medium

Abstract

The present application relates to a battery parameter identification method, device, computer equipment and storage medium. The method comprises: based on an equivalent circuit model of a current battery, obtaining a constraint parameter identification model corresponding to the equivalent circuit model, wherein the constraint parameter identification model is a parameter identification model including constraint conditions; based on the constraint parameter identification model and timing operating condition data, obtaining a first characteristic parameter of the current battery; and based on the first characteristic parameter input into a neural network model, obtaining a parameter identification result. The use of this method can improve the accuracy of battery parameter identification.

Description

Technical Field

[0001] The present application relates to the field of battery management technology, and in particular to a battery parameter identification method, device, computer equipment and storage medium. Background Art

[0002] With the development of new energy electric vehicle technology, accurate estimation of battery SOC (State of Charge) is becoming increasingly important for the use and safety of electric vehicles. Currently, SOC estimation is usually corrected and estimated through the ampere-hour integration method, which involves parameter identification of the equivalent circuit model.

[0003] Through offline parameter identification, due to the complexity of the vehicle operating environment and the uncertainty of battery cell aging, it is easy to cause the identification parameters to be inconsistent with reality, while the current parameter identification of equivalent circuit models still has the problem of poor accuracy. Summary of the invention

[0004] Based on this, it is necessary to provide a battery parameter identification method, device, computer equipment and storage medium that can improve the accuracy of battery parameter identification in order to solve the above technical problems.

[0005] In a first aspect, the present application provides a battery parameter identification method, the battery parameter identification method comprising:

[0006] Based on an equivalent circuit model of the current battery, a constraint parameter identification model corresponding to the equivalent circuit model is obtained, wherein the constraint parameter identification model is a parameter identification model including constraint conditions;

[0007] Based on the constraint parameter identification model and the time sequence operating condition data, obtaining a first characteristic parameter of the current battery;

[0008] Based on the first characteristic parameter being input into the neural network model, a parameter identification result is obtained.

[0009] In one embodiment, the equivalent circuit model includes: DC internal resistance, polarization internal resistance and polarization capacitance.

[0010] In one embodiment, obtaining a constraint parameter identification model corresponding to the equivalent circuit model based on the equivalent circuit model includes:

[0011] Based on the equivalent circuit model, constructing a transfer function corresponding to the equivalent circuit model;

[0012] Performing bilinear transformation on the transfer function to obtain a discretized equation;

[0013] Based on the discretized equation, determining a parameter identification model of the equivalent circuit model;

[0014] Boundary constraints are imposed on the parameters of the parameter identification model to obtain a constrained parameter identification model.

[0015] In one embodiment, the discretization equation includes a DC internal resistance coefficient, a polarization internal resistance coefficient, and a polarization time coefficient, and the parameter identification model for determining the equivalent circuit model based on the discretization equation includes:

[0016] Determining identification model parameters of the equivalent circuit model based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient; the identification model parameters include a historical voltage influence vector, a current current influence vector, a historical current influence vector and an open circuit voltage contribution vector;

[0017] Based on the identification model parameters, a parameter identification model of the equivalent circuit model is determined.

[0018] In one embodiment, performing boundary constraints on the parameters of the parameter identification model to obtain the constrained parameter identification model includes:

[0019] Determining a constraint parameter range of the historical voltage influence vector based on the polarization time;

[0020] Based on the DC internal resistance, determining a first constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector;

[0021] Based on the polarization internal resistance, determining a second constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector;

[0022] Based on the constraint parameter range, the first constraint relationship and the second constraint relationship, boundary constraints are performed on the parameters of the parameter identification model to obtain a constraint parameter identification model.

[0023] In one embodiment, determining the constraint parameter range of the historical voltage influence vector based on the polarization time includes:

[0024] Determining a minimum value of the historical voltage influence vector based on the minimum value of the polarization time;

[0025] Based on the maximum value of the polarization time, a maximum value of the historical voltage influence vector is determined.

[0026] In one embodiment, determining the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the DC internal resistance includes:

[0027] Based on the relationship between the DC internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the DC internal resistance, a first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

[0028] In one embodiment, determining the second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the polarization internal resistance includes:

[0029] Based on the relationship between the polarization internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the polarization internal resistance, a second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

[0030] In one embodiment, the sequential operating condition data includes voltage data and current data at multiple different temperatures and charge states, and obtaining the first characteristic parameter of the current battery based on the parameter identification model and the sequential operating condition data includes:

[0031] Based on the parameter identification model and the time series operating condition data, sequentially calculating the preliminary characteristic parameters of the current battery at each moment;

[0032] Performing stability verification on the preliminary characteristic parameters at each moment;

[0033] Based on the preliminary characteristic parameters at multiple moments verified by stability, the first characteristic parameters of the current battery are obtained.

[0034] In one embodiment, the step of inputting the first characteristic parameter into a neural network model to obtain a parameter identification result includes:

[0035] Based on the first characteristic parameter, calculating the terminal voltage of the current battery (the battery terminal voltage calculation method is to substitute the characteristic parameter output by the neural network into the equivalent circuit, and calculate the terminal voltage according to the historical state of charge, current, and temperature data);

[0036] Calculating a residual of the neural network model based on a difference between a measured voltage and the terminal voltage;

[0037] Based on the residual, optimizing the neural network model;

[0038] Based on the first characteristic parameter and the optimized neural network model, a parameter identification result is obtained.

[0039] In one embodiment, the calculating the current terminal voltage of the battery based on the first characteristic parameter includes:

[0040] Inputting the first characteristic parameter into the neural network model to obtain a second characteristic parameter;

[0041] Substituting the second characteristic parameter and the monitoring operating condition data into the equivalent circuit model, the terminal voltage is obtained; the monitoring operating condition data includes historical state of charge, current and temperature.

[0042] In one embodiment, the neural network model is optimized based on the residual; and the parameter identification result is obtained based on the first characteristic parameter and the optimized neural network model, including:

[0043] When the residual is greater than a preset residual threshold, the weight parameters of the neural network model are updated based on the residual, and the residual of the neural network model is recalculated;

[0044] When the residual is less than or equal to a preset residual threshold, a weight operation is performed on the first characteristic parameter based on the optimized neural network model to obtain an optimized characteristic parameter; and a parameter identification result is obtained based on the optimized characteristic parameter.

[0045] In a second aspect, the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above method when executing the computer program.

[0046] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method described above when executed by a processor.

[0047] In a fourth aspect, the present application provides a computer program product, which implements the method described above when the computer program is executed by a processor.

[0048] The above-mentioned battery parameter identification method, device, computer equipment and storage medium obtain a constrained parameter identification model corresponding to the equivalent circuit model based on the equivalent circuit model of the current battery, and the constrained parameter identification model is a parameter identification model including constraint conditions; based on the constrained parameter identification model and timing operating condition data, the first characteristic parameter of the current battery is obtained; based on the first characteristic parameter, it is input into the neural network model to obtain a parameter identification result, and the first characteristic parameter that meets the operating conditions and satisfies the constraint conditions can be preliminarily obtained through the constrained parameter identification model and the timing operating condition data, which can not only obtain more accurate battery characteristic parameters, but also input the characteristic parameters identified through constraints into the neural network model for weight calculation optimization, and can also avoid the neural network model from obtaining locally optimal parameter results, thereby achieving the effect of improving the accuracy of battery parameter identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A diagram showing an application environment of a battery parameter identification method in one embodiment;

[0050] Figure 2 is a schematic flow chart of a battery parameter identification method in one embodiment;

[0051] Figure 3 is a circuit diagram of an equivalent circuit model in an embodiment;

[0052] Figure 4 is a flow chart of a battery parameter identification method according to another embodiment;

[0053] Figure 5 A schematic diagram of the structure of a neural network in one embodiment;

[0054] Figure 6 is a structural block diagram of a battery parameter identification device in one embodiment;

[0055] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0057] The battery parameter identification method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 is connected to the battery detection device 104 in communication, and the terminal 102 can be integrated with a data storage system. Based on the equivalent circuit model of the current battery, the terminal 102 obtains a constraint parameter identification model corresponding to the equivalent circuit model, and the constraint parameter identification model is a parameter identification model containing constraint conditions; based on the constraint parameter identification model and the timing condition data, the first characteristic parameter of the current battery is obtained; based on the first characteristic parameter, it is input into the neural network model to obtain a parameter identification result. Among them, the terminal 102 can be a control chip such as MCU (Microcontroller Unit), FPGA (Field Programmable Gate Array), SoC (System on Chip, System on Chip). The battery detection device 104 can be used to perform voltage detection, current detection, temperature detection, etc. on the battery pack and / or the monomers in the battery pack.

[0058] In one embodiment, Figure 2 As shown, a battery parameter identification method is provided, which is applied to Figure 1 The terminal 102 in the example is used as an example to illustrate, and the following steps are included:

[0059] Step S100, based on the equivalent circuit model of the current battery, obtaining a constraint parameter identification model corresponding to the equivalent circuit model.

[0060] Among them, the equivalent circuit model is a model used to simplify the analysis of complex circuits. It can use a combination of ideal circuit elements such as resistors, capacitors, and inductors to approximate the actual circuit, thereby facilitating calculation and analysis.

[0061] The constrained parameter identification model is a parameter identification model that includes constraint conditions. The parameter identification model can be first established based on the equivalent circuit model, and then the parameters of the parameter identification model are conditionally constrained to obtain the constrained parameter identification model. The constraint conditions can be set according to the physical and chemical properties of the battery, such as temperature dependence, current density limitation, etc., so that the first characteristic parameter obtained is reasonable in a physical sense.

[0062] Based on the equivalent circuit model of the current battery, a corresponding set of mathematical equations can be established according to the ideal circuit elements in the equivalent circuit model as a parameter identification model. The parameter identification model can characterize the theoretical relationship between the battery terminal voltage and the ideal circuit elements.

[0063] In this embodiment, conditional constraints are imposed on the parameters of the parameter identification model themselves, and then the constrained parameter identification model with the constraint conditions is used to perform parameter identification to obtain the first characteristic parameters. Compared with the parameter distortion that may be caused by first identifying the parameters based on the parameter identification model and then constraining the characteristic parameters, a more accurate and appropriate first characteristic parameter can be obtained.

[0064] Step S200, obtaining the first characteristic parameter of the current battery based on the constraint parameter identification model and the time sequence operating condition data.

[0065] Among them, the sequential operating condition data can be a data record of the battery changing over time under different operating conditions, and further, it can be a data record of the battery changing over time under different charge states and different temperatures. Exemplarily, the sequential operating condition data may include data such as charge and discharge current and terminal voltage. The sequential operating condition data can be data obtained by testing under controlled conditions, or it can be data collected during the operation of the device where the battery is located.

[0066] Before obtaining the first characteristic parameter through the time series operating condition data, the time series operating condition data may be preprocessed, exemplarily by methods including denoising, filling missing values, time series alignment, etc., to improve data quality.

[0067] Based on the constraint parameter identification model and the sequential operating condition data, the constraint parameter identification model can be fitted by the sequential operating condition data, so that when the obtained characteristic parameters are applied to the constraint parameter identification model, the output result is similar to or the same as the sequential operating condition data. The first characteristic parameter can be used to characterize the key parameters of the battery performance at different charge states and / or different temperatures. For example, the key parameter can be one or more parameters such as DC internal resistance, polarization internal resistance, polarization capacitance, open circuit voltage, etc.

[0068] Step S300, inputting the first characteristic parameter into a neural network model to obtain a parameter identification result.

[0069] The first characteristic parameter is input into the neural network model by taking the first characteristic parameter as an initial value and optimizing the first characteristic parameter by using the weight operation of the neural network model to obtain a parameter identification result.

[0070] The neural network model may be a feedforward neural network, a convolutional neural network, a recurrent neural network, a long short-term memory network, or other existing neural network models, which is not limited in this embodiment.

[0071] A battery parameter identification method provided in this embodiment obtains a constrained parameter identification model corresponding to the equivalent circuit model based on an equivalent circuit model of the current battery, wherein the constrained parameter identification model is a parameter identification model including constraint conditions; obtains a first characteristic parameter of the current battery based on the constrained parameter identification model and timing operating condition data; obtains a parameter identification result based on the first characteristic parameter input into a neural network model, and can preliminarily obtain a first characteristic parameter that meets the operating condition and satisfies the constraint conditions through the constrained parameter identification model and the timing operating condition data, which can not only obtain more accurate battery characteristic parameters, but also input the characteristic parameters identified through the constraints into the neural network model for weight calculation optimization, and can also avoid the neural network model from obtaining a locally optimal parameter result, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0072] like Figure 3 As shown, in one embodiment, the equivalent circuit model includes: DC internal resistance, polarization internal resistance and polarization capacitance.

[0073] Among them, the DC internal resistance represents the inherent resistance inside the battery, which can be recorded as R0, the polarization internal resistance is related to the internal chemical reaction of the battery, which can be recorded as R1 or Rp, and the polarization capacitance can represent the internal chemical reaction speed of the battery, which can be recorded as C1 or Cp. Ut is the terminal voltage, and Uoc is the open circuit voltage.

[0074] Furthermore, the constraint conditions may be set according to the physical meanings of the DC internal resistance, polarization internal resistance, and polarization capacitance.

[0075] A battery parameter identification method provided in this embodiment can more accurately reflect battery characteristics by forming an equivalent circuit model based on DC internal resistance, polarization internal resistance and polarization capacitance, thereby enhancing the adaptability of the model to different working conditions and achieving the effect of improving the accuracy of battery parameter identification.

[0076] In one embodiment, obtaining a constraint parameter identification model corresponding to the equivalent circuit model based on the equivalent circuit model includes:

[0077] Based on the equivalent circuit model, constructing a transfer function corresponding to the equivalent circuit model;

[0078] Performing bilinear transformation on the transfer function to obtain a discretized equation;

[0079] Based on the discretized equation, determining a parameter identification model of the equivalent circuit model;

[0080] Boundary constraints are imposed on the parameters of the parameter identification model to obtain a constrained parameter identification model.

[0081] The transfer function is used to describe the behavior of the system in the frequency domain. The transfer function can be obtained by performing Laplace transform on the equivalent circuit model.

[0082] Bilinear transformation is used to convert a continuous-time system into a discrete-time system by mapping the s-plane to the z-plane to obtain a discretized equation to maintain the stability and frequency response characteristics of the system. The discretized equation is an equation that describes the behavior of the system in the discrete time domain.

[0083] Based on the discretized equation, the parameter identification model of the equivalent circuit model is determined, and a mathematical model can be constructed based on the discretized equation to estimate the parameters in the equivalent circuit model. Specifically, coefficients related to battery parameters can be extracted from the discretized equation to form a parameter identification model.

[0084] Boundary constraints are imposed on the parameters of the parameter identification model, which may be to restrict the minimum and maximum values ​​of the parameters in the parameter identification model, or to restrict the magnitude relationship between at least two parameters, thereby obtaining a constrained parameter identification model.

[0085] A battery parameter identification method provided in this embodiment obtains a parameter identification model by constructing a transfer function of an equivalent circuit model and performing a bilinear transformation, and then performs boundary constraints on the parameters to obtain a constrained parameter identification model. A relatively accurate parameter identification model can be constructed, and when the parameter identification model contains constraints, more accurate identification parameters can be obtained, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0086] In one embodiment, the discretization equation includes a DC internal resistance coefficient, a polarization internal resistance coefficient, and a polarization time coefficient, and the parameter identification model for determining the equivalent circuit model based on the discretization equation includes:

[0087] Determining identification model parameters of the equivalent circuit model based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient; the identification model parameters include a historical voltage influence vector, a current current influence vector, a historical current influence vector and an open circuit voltage contribution vector;

[0088] Based on the identification model parameters, a parameter identification model of the equivalent circuit model is determined.

[0089] Among them, the DC internal resistance coefficient is a variable related to the DC internal resistance, the polarization internal resistance coefficient is a variable related to the polarization internal resistance, and the polarization time coefficient is a variable related to the time quantity, wherein the time quantity is the product of the polarization resistance and the polarization capacitance.

[0090] The historical voltage influence vector is used to represent the influence of the polarization voltage at the previous moment on the current terminal voltage, the current current influence vector is used to represent the influence of the current current on the current terminal voltage, the historical current influence vector is used to represent the influence of the current at the previous moment on the current terminal voltage, and the open circuit voltage contribution vector is used to represent the contribution of the open circuit voltage to the current terminal voltage.

[0091] By transforming and z-inverse transforming the discretized equation, the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient in the discretized equation can be respectively composed of the historical voltage influence vector, the current current influence vector, the historical current influence vector and the open circuit voltage contribution vector, thereby determining the parameter identification model of the final equivalent circuit model.

[0092] In one specific embodiment, the transfer function is:

[0093]

[0094] Among them, Ut is the terminal voltage, Uoc is the open circuit voltage, I is the current, R0 is the DC internal resistance, R1 is the polarization internal resistance, τ is the polarization time, and s is the time.

[0095] By performing a bilinear transformation on the transfer function, we can obtain the discretized equation:

[0096] and

[0097] Among them, A0 is the DC internal resistance coefficient, A1 is the polarization internal resistance coefficient, and B1 is the polarization time coefficient.

[0098] Further, based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient, the identification model parameters of the equivalent circuit model are determined by transforming the discretized equation to obtain:

[0099]

[0100] Since Uoc changes slightly between the previous and next sampling points, it can be regarded as unchanged, so the z-transform can be obtained:

[0101]

[0102] Based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient, the identification model parameters of the equivalent circuit model are determined, which can be As the open circuit voltage contribution vector θ4, As the historical voltage influence vector θ1, As the current influence vector θ2, As the historical current influence vector θ3, the identification model parameters of the equivalent circuit model are obtained;

[0103] Furthermore, based on the identification model parameters, the parameter identification model of the equivalent circuit model is determined, and the identification equation can be obtained according to the above parameters, that is:

[0104]

[0105] A battery parameter identification method provided in this embodiment determines the identification model parameters of the equivalent circuit model by discretizing the DC internal resistance coefficient, polarization internal resistance coefficient and polarization time coefficient in the equation, thereby further constructing a parameter identification model. This can effectively improve the accuracy of parameter identification and ensure that the identified parameters have physical meaning and rationality, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0106] In one embodiment, performing boundary constraints on the parameters of the parameter identification model to obtain the constrained parameter identification model includes:

[0107] Determining a constraint parameter range of the historical voltage influence vector based on the polarization time;

[0108] Based on the DC internal resistance, determining a first constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector;

[0109] Based on the polarization internal resistance, determining a second constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector;

[0110] Based on the constraint parameter range, the first constraint relationship and the second constraint relationship, boundary constraints are performed on the parameters of the parameter identification model to obtain a constraint parameter identification model.

[0111] Wherein, the polarization time may be the product of the polarization capacitance and the polarization resistance, and the historical voltage influence vector may represent the influence of the polarization voltage at the previous moment on the current terminal voltage. Based on the polarization time, the constraint parameter range of the historical voltage influence vector is determined, and the range of the historical voltage influence vector may be determined accordingly based on the relationship between the polarization time and the historical voltage influence vector, and the range of the polarization time, to obtain the constraint parameter range.

[0112] The DC internal resistance is used to represent the ohmic internal resistance of the battery. It can be understood that the DC internal resistance value can be obtained through the historical voltage influence vector, the current current influence vector and the historical voltage influence vector. Based on the DC internal resistance, the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined. The historical voltage influence vector, the current current influence vector and the historical voltage influence vector can be restricted within the first constraint relationship according to the range of the DC internal resistance.

[0113] The polarization internal resistance is used to represent the ohmic internal resistance of the battery. It can be understood that the polarization internal resistance value can be obtained through the historical voltage influence vector, the current current influence vector and the historical voltage influence vector. Based on the polarization internal resistance, the second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined. The historical voltage influence vector, the current current influence vector and the historical voltage influence vector can be restricted within the second constraint relationship according to the range of the polarization internal resistance.

[0114] Based on the constraint parameter range, the first constraint relationship and the second constraint relationship, boundary constraints are imposed on the parameters of the parameter identification model. The constraint parameter range, the first constraint relationship and the second constraint relationship may be added to the parameter identification model. That is, in the parameter identification process, the first characteristic parameter obtained not only needs to satisfy the parameter identification model and the timing operating condition data, but also needs to satisfy the constraint parameter range, the first constraint relationship and the second constraint relationship, so as to realize the identification of the first characteristic parameter.

[0115] A battery parameter identification method provided in this embodiment determines the constraint parameter range, the first constraint relationship and the second constraint relationship of the historical voltage influence vector, the current current influence vector and the historical current influence vector through polarization time, DC internal resistance and polarization internal resistance, and substitutes them into the parameter identification model to obtain the constraint parameter identification model, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0116] In one embodiment, determining the constraint parameter range of the historical voltage influence vector based on the polarization time includes:

[0117] Determining a minimum value of the historical voltage influence vector based on the minimum value of the polarization time;

[0118] Based on the maximum value of the polarization time, a maximum value of the historical voltage influence vector is determined.

[0119] Wherein, determining the constraint parameter range of the historical voltage influence vector based on the polarization time may be to determine the maximum value and the minimum value of the historical voltage influence vector respectively according to the maximum value and the minimum value of the polarization time.

[0120] In a specific embodiment, That is the historical voltage influence vector θ1, that is, B1=-θ1. According to the relationship between B1 and polarization time τ1, we can get:

[0121]

[0122] You will get:

[0123]

[0124] It can be seen that the historical voltage influence vector θ1 is a monotonic function of the polarization time τ1. min <τ<τ max When θ 1,min <θ1<θ 1,max .

[0125] A battery parameter identification method provided in this embodiment can ensure that the historical voltage influence vector is within a reasonable range by determining the maximum and minimum values ​​of the historical voltage influence vector according to the maximum and minimum values ​​of the polarization time, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0126] In one embodiment, determining the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the DC internal resistance includes:

[0127] Based on the relationship between the DC internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the DC internal resistance, a first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

[0128] The DC internal resistance represents the ohmic internal resistance of the battery.

[0129] Since the identification model parameters of the equivalent circuit model can be determined based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient, the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient can be represented by the identification model parameters accordingly.

[0130] In a specific embodiment, the following relationship can be obtained:

[0131]

[0132] Among them, θ1 is the historical voltage influence vector, θ2 is the current current influence vector, θ3 is the historical current influence vector, θ4 is the open circuit voltage contribution vector, A0 is the DC internal resistance coefficient, A1 is the polarization internal resistance coefficient, and B1 is the polarization time coefficient.

[0133] From this we can get:

[0134]

[0135] By setting the minimum value of the DC internal resistance, the identification value of the DC internal resistance can be made to be at least greater than the minimum value. Since A0=R0, when R0 has a minimum value R 0,min When , we can get:

[0136]

[0137] After transformation, we can get:

[0138]

[0139] in, is the identification value of the DC internal resistance. Thus, the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector can be obtained.

[0140] A battery parameter identification method provided in this embodiment determines a first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the relationship between the DC internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the DC internal resistance. This can avoid the first characteristic parameter obtained by identification from exceeding the normal range of the DC internal resistance, thereby improving the accuracy and reliability of parameter identification and achieving the effect of improving the accuracy of battery parameter identification.

[0141] In one embodiment, determining the second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the polarization internal resistance includes:

[0142] Based on the relationship between the polarization internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the polarization internal resistance, a second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

[0143] Among them, since the identification model parameters of the equivalent circuit model can be determined based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient, correspondingly, the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient can be represented by the identification model parameters.

[0144] In a specific embodiment, the following relationship is known:

[0145]

[0146] Among them, θ1 is the historical voltage influence vector, θ2 is the current current influence vector, θ3 is the historical current influence vector, θ4 is the open circuit voltage contribution vector, A0 is the DC internal resistance coefficient, A1 is the polarization internal resistance coefficient, and B1 is the polarization time coefficient. It can be obtained:

[0147]

[0148] By setting the minimum value of the polarization internal resistance, the identification value of the polarization internal resistance can be made to be at least greater than the minimum value. 1,min When , we have:

[0149]

[0150] After transformation, we can get:

[0151]

[0152] When the identification value of polarization internal resistance The minimum value of R 1,min When , we have:

[0153]

[0154] Combination By transformation, we can get:

[0155]

[0156] Then we can get:

[0157]

[0158] Thus, a second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector can be obtained. The second constraint relationship includes:

[0159]

[0160] In addition, since the polarization capacitance is the ratio of the polarization time to the polarization internal resistance, the polarization time can be bounded based on the cell test, so the boundary constraints of the polarization capacitance can also be obtained.

[0161] A battery parameter identification method provided in this embodiment determines the second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the relationship between the polarization internal resistance and the historical voltage influence vector, the current current influence vector and the historical voltage influence vector, and the minimum value of the polarization internal resistance. This can avoid the first characteristic parameter obtained by identification from exceeding the normal range of the polarization internal resistance, thereby improving the accuracy and reliability of parameter identification and achieving the effect of improving the accuracy of battery parameter identification.

[0162] In one embodiment, obtaining the first characteristic parameter of the current battery based on the parameter identification model and the time sequence operating condition data includes:

[0163] Based on the parameter identification model and the time series operating condition data, sequentially calculating the preliminary characteristic parameters of the current battery at each moment;

[0164] Performing stability verification on the preliminary characteristic parameters at each moment;

[0165] Based on the preliminary characteristic parameters at multiple moments verified by stability, the first characteristic parameters of the current battery are obtained.

[0166] The time sequence operating condition data includes voltage data and current data at multiple different temperatures and states of charge. The temperature can be the ambient temperature when the battery is working, and the state of charge (SOC) can be understood as the charging level of the battery.

[0167] Based on the parameter identification model and the time series operating condition data, the preliminary characteristic parameters of the current battery at each moment are calculated in sequence. The terminal voltage of the battery can be predicted using the parameter identification model according to the voltage and current at the current moment, and the error between the predicted value and the actual measured value is calculated. The parameters in the parameter identification model are adjusted based on the optimization algorithm to minimize the error. It is understandable that the optimization algorithm can be implemented by one or more methods such as the gradient descent method and the least squares method.

[0168] The stability verification of the preliminary characteristic parameters at each moment may be performed by taking the current moment as the end moment, obtaining the change trend of the preliminary characteristic parameters within the historical time range according to the preset time length, judging whether the change of the preliminary characteristic parameters is within the stable threshold range, ensuring that there is no sudden change or abnormal fluctuation in the parameters over time, or that the mean and standard deviation of the parameters are within a reasonable range, thereby completing the stability verification.

[0169] The first characteristic parameter may be a battery characteristic parameter at different states of charge and different temperatures. The first characteristic parameter of the current battery is obtained based on the preliminary characteristic parameters at multiple moments verified by stability, and the first characteristic parameter of the current battery is obtained by combining multiple preliminary characteristic parameters verified by stability.

[0170] When the stability verification fails, new preliminary characteristic parameters may be obtained through an optimization algorithm and verified again, and after passing the verification, used to determine the first characteristic parameters.

[0171] This embodiment provides a battery parameter identification method, which calculates the preliminary characteristic parameters at each moment in turn based on the parameter identification model and the time series operating condition data, and performs stability verification. Finally, the first characteristic parameters of the current battery are obtained by aggregating the qualified parameters that have passed the stability verification. This method can not only improve the accuracy of parameter identification, but also ensure the stability and reliability of the parameters under different operating conditions, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0172] In one embodiment, in addition to obtaining the first characteristic parameter by using the optimization algorithm, the time sequence operating condition data and the parameter identification model, the parameter is further optimized by using a neural network, and the parameter identification result obtained by inputting the first characteristic parameter into the neural network model includes:

[0173] Calculating the current terminal voltage of the battery based on the first characteristic parameter;

[0174] Calculating a residual of the neural network model based on a difference between a measured voltage and the terminal voltage;

[0175] Based on the residual, optimizing the neural network model;

[0176] Based on the first characteristic parameter and the optimized neural network model, a parameter identification result is obtained.

[0177] The calculation of the current terminal voltage of the battery based on the first characteristic parameter may be performed by substituting the first characteristic parameter into an equivalent circuit model to calculate the current terminal voltage of the battery under the current first characteristic parameter.

[0178] In this embodiment, the first characteristic parameter may be used as the initial value of the neural network model, and the residual of the neural network model may be calculated based on the difference between the measured voltage and the terminal voltage. The difference may be further converted to obtain the residual of the neural network model.

[0179] It can be understood that initial weights are set in the neural network model. Based on the residual, the neural network model is optimized. The weights of the neural network model can be optimized using a loss function, so that after the optimized neural network model processes the first characteristic parameter, the residual can obtain the battery characteristic parameter that meets the requirements as the parameter identification result.

[0180] A battery parameter identification method provided in this embodiment calculates the terminal voltage of the battery by inputting a first characteristic parameter into a neural network model, calculates the residual based on the difference between the measured voltage and the terminal voltage, and minimizes the residual by optimizing the neural network model. Finally, an optimized parameter identification result is obtained, which can improve the accuracy of parameter identification and ensure the stability and reliability of the parameters under different working conditions, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0181] In one embodiment, the calculating the current terminal voltage of the battery based on the first characteristic parameter includes:

[0182] Inputting the first characteristic parameter into the neural network model to obtain a second characteristic parameter;

[0183] Substituting the second characteristic parameter and the monitoring operating condition data into the equivalent circuit model, the terminal voltage is obtained; the monitoring operating condition data includes historical state of charge, current and temperature.

[0184] Exemplarily, the first characteristic parameter can be processed according to the current architecture and weight distribution of the neural network model to obtain the second characteristic parameter, which is the characteristic parameter obtained by calculating the current weight of the neural network model with the first characteristic parameter. By substituting the second characteristic parameter and the monitoring condition data into the equivalent circuit model for calculation, the theoretical value of the terminal voltage under the second characteristic parameter and the monitoring condition data can be obtained, and then the difference between the theoretical value and the measured value can be used to determine whether there is a large error in the weight of the current neural network model.

[0185] A battery parameter identification method provided in this embodiment, by substituting the second characteristic parameter and the monitoring condition data into the equivalent circuit model to obtain the terminal voltage, can improve the accuracy of parameter identification, and can also ensure the stability and reliability of the parameters under different conditions, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0186] In one embodiment, the neural network model is optimized based on the residual; and the parameter identification result is obtained based on the first characteristic parameter and the optimized neural network model, including:

[0187] When the residual is greater than a preset residual threshold, the weight parameters of the neural network model are updated based on the residual, and the residual of the neural network model is recalculated;

[0188] When the residual is less than or equal to a preset residual threshold, a weight operation is performed on the first characteristic parameter based on the optimized neural network model to obtain an optimized characteristic parameter; and a parameter identification result is obtained based on the optimized characteristic parameter.

[0189] In this embodiment, whether to continuously optimize the neural network model can be determined based on whether the residual is less than or equal to a preset residual threshold. When the residual is greater than the preset residual threshold, it means that the weight of the current neural network model has not yet reached the optimal state, and the weight parameters need to be updated using the loss function and iteratively optimized. When the residual is less than or equal to the preset residual threshold, it means that the weight parameters of the current neural network model have met expectations, and the first characteristic parameter can be weighted by the current neural network model to obtain the optimized characteristic parameter.

[0190] Based on the optimized characteristic parameters, a parameter identification result is obtained, and the optimized characteristic parameters may be used as the parameter identification result and information is returned.

[0191] A battery parameter identification method provided in this embodiment can ensure that the weight parameters of the neural network model meet the expected values ​​by iteratively optimizing the neural network model based on whether the residual is less than or equal to a preset residual threshold, thereby improving the accuracy of parameter identification, and can also ensure the stability and reliability of the parameters under different working conditions, thereby achieving the effect of improving the accuracy of battery parameter identification.

[0192] In order to more clearly illustrate the technical solution of the present application, the present application also provides a detailed embodiment.

[0193] In one embodiment, Figure 4 As shown, a battery parameter identification method is provided, comprising:

[0194] Step S1, performing constrained parameter identification according to battery cell operating condition data.

[0195] According to the RC equivalent circuit model, a parameter identification formula with constraints is established. According to the circuit model, the expression can be obtained:

[0196]

[0197] Among them, Ut is the terminal voltage, Uoc is the open circuit voltage, I is the current, R0 is the DC internal resistance, R1 is the polarization internal resistance, τ is the polarization time, and s is the time.

[0198] By performing a bilinear transformation on the transfer function, we can obtain the discretized equation:

[0199] and

[0200] Among them, A0 is the DC internal resistance coefficient, A1 is the polarization internal resistance coefficient, and B1 is the polarization time coefficient.

[0201] The DC internal resistance coefficient A0 is a variable related to the DC internal resistance R0, the polarization internal resistance coefficient A1 is a variable related to the polarization internal resistance R1, and the polarization time coefficient B1 is a variable related to the time quantity, wherein the time quantity τ is the product of the polarization resistance R1 and the polarization capacitance C1, that is, τ1=R1×C1.

[0202] Based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient, the identification model parameters of the equivalent circuit model are determined by transforming the discretization equation to obtain:

[0203]

[0204] Since Uoc changes slightly between the previous and next sampling points, it can be regarded as unchanged, so the z-transform can be obtained:

[0205]

[0206] Based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient, the identification model parameters of the equivalent circuit model are determined, which can be As the open circuit voltage contribution vector θ4, As the historical voltage influence vector θ1, As the current influence vector θ2, As the historical current influence vector θ3, the identification model parameters of the equivalent circuit model are obtained;

[0207] Furthermore, based on the identification model parameters, the parameter identification model of the equivalent circuit model is determined, and the identification equation can be obtained according to the above parameters, that is:

[0208]

[0209] From this we can get:

[0210]

[0211] Therefore, the RC parameters of the battery cell can be solved according to the least squares method to achieve convergence of R0, R1, and C1 parameters.

[0212] In order to prevent the RC parameter identification from falling into the local optimum and speed up the identification process, the parameters need to be constrained. The following is the derivation and establishment of the constraint formula.

[0213] Combining the above formula, we can get:

[0214]

[0215] Among them, θ1 is the historical voltage influence vector, θ2 is the current current influence vector, θ3 is the historical current influence vector, θ4 is the open circuit voltage contribution vector, A0 is the DC internal resistance coefficient, A1 is the polarization internal resistance coefficient, and B1 is the polarization time coefficient.

[0216] From this we can get:

[0217]

[0218] According to the physical requirements of the A and B model parameters, it can be seen that A and B must be > 0.

[0219] In order to ensure that the sign of the solver convergence result is always positive, a minimum threshold needs to be given, so setting τ1 needs to satisfy τ min <τ<τ max (τ1 is the polarization time), τ min and τ max It can be obtained by battery cell testing.

[0220] That is the historical voltage influence vector θ1, that is, B1=-θ1. According to the relationship between B1 and polarization time τ1, we can get:

[0221]

[0222] You will get:

[0223]

[0224] It can be seen that the historical voltage influence vector θ1 is a monotonic function of the polarization time τ1. min <τ<τ max When θ 1,min <θ1<θ 1,max .

[0225] By setting the minimum value of the DC internal resistance, the identification value of the DC internal resistance can be made to be at least greater than the minimum value. Since A0=R0, when R0 has a minimum value R 0,min When , we can get:

[0226]

[0227] After transformation, we can get:

[0228]

[0229] in, is the identification value of the DC internal resistance. Thus, the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector can be obtained.

[0230] By setting the minimum value of the polarization internal resistance, the identification value of the polarization internal resistance can be made to be at least greater than the minimum value. 1,min When , we have:

[0231]

[0232] After transformation, we can get:

[0233]

[0234] When the identification value of polarization internal resistance The minimum value of R 1,min When , we have:

[0235]

[0236] Combination By transformation, we can get:

[0237]

[0238] Then we can get:

[0239]

[0240] Thus, a second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector can be obtained. The second constraint relationship includes:

[0241] Therefore, according to the above derivation, the boundary constraint can be summarized as follows:

[0242]

[0243] in, include:

[0244]

[0245] In addition, since the polarization capacitance C1 is the ratio of the polarization time τ to the polarization internal resistance R1, the polarization time τ can be bounded according to the cell test, so the boundary constraint of the polarization capacitance C1 can also be obtained. Thus, the cell parameters R0, R1, and C1 all have boundary constraints.

[0246] Step S2, according to the voltage and current data of the battery cell under different temperatures and different SOCs under the test conditions, substitute the constrained identification formula of S1 to perform RC parameter identification and obtain R0 k , R1 k , C1 k , where k is time. To verify the stability of the parameters, F(R0, k) = |R0 k -R0 k-n |, F(R1, k) = |R1 k -R1 k-n |, F(C1, k) = |C1 k -C1 k-n |, determine whether F≤Q (stability threshold parameter) and last for x time (time threshold parameter). If the output parameters include the R00, R10, and C10 tables of the temperature and SOC dimensions, execute S3; otherwise, continue with identification.

[0247] Step S3, use the parameters R0, R1, and C1 output in S2 as the initial value of the neural network, and QZ0 as the initial value of the weight of the neural network. After the neural network operation, output R0bp, R1bp, and C1bp. Among them, R0bp is the identification of the DC internal resistance, R1bp is the identification of the polarization internal resistance, and C1bp is the identification of the polarization capacitance. Figure 5 As shown in the figure, there are 3 input RC parameters, 3 neural network input nodes, 7 first hidden layer nodes, 9 second hidden layer nodes, 7 third hidden layer nodes, and 3 output nodes, which are the identified model parameters.

[0248] Substitute R0bp, R1bp, and C1bp into the equivalent circuit model, and calculate the terminal voltage f(R0bp, R1bp, C1bp, Temp, SOC) based on the equivalent circuit model and the current temperature and SOC value. Calculate the residual based on the measured voltage Y and the terminal voltage , n is the time.

[0249] The stochastic gradient algorithm is used to substitute the residual into the neural network to update the weight QZ, and then the RC is weighted again to obtain new RC parameters.

[0250] As follows:

[0251]

[0252]

[0253] in, represents the loss function, J(E) is equivalent to the gradient of parameter E (residual), and n represents time.

[0254] Determine whether E converges: Stop the neural network training by judging whether the residual E ≤ G (residual threshold). If it is satisfied, output the RC parameters; if it does not converge, substitute E into the stochastic gradient algorithm to update the internal weights of the neural network and continue training.

[0255] Through training, the RC parameters of the battery cells at different temperatures and different SOCs can be obtained.

[0256] A battery parameter identification method provided in this embodiment adopts an RC equivalent circuit, based on the least squares method, and derives RC boundary constraints. According to the battery cell data, RC parameter identification is performed to obtain pre-processed RC parameters. Then, the RC parameters obtained by pre-processing are used as the initial values ​​of the neural network, and the RC parameters processed by the weights according to the vehicle's historical voltage, current, temperature data and the output of the neural network are substituted into the equivalent circuit. The difference between the terminal voltage calculated by the equivalent circuit and the actual test voltage is used as the residual, and the residual is substituted into the neural network through the stochastic gradient algorithm to update the weights, and the RC parameters are output again. Through continuous training, accurate battery equivalent circuit model parameters are obtained; by adding constraints to the identification formula, the parameter identification process can be prevented from not converging; the RC parameters after constraint identification are used as the initial values ​​of the neural network to improve training efficiency and accuracy; the RC parameters output by the neural network are substituted into the equivalent circuit model to calculate the terminal voltage, and the weights of the neural network are updated according to the residual calculated by the measured voltage. It can ensure that the training parameters are within the physical constraints, avoid the numerical value from falling into the local optimum, and provide identification accuracy at the same time.

[0257] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0258] Based on the same inventive concept, the embodiment of the present application also provides a battery parameter identification device for implementing the battery parameter identification method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more battery parameter identification device embodiments provided below can refer to the limitations of the battery parameter identification method above, and will not be repeated here.

[0259] In one embodiment, Figure 6 As shown, a battery parameter identification device is provided, comprising: a model building module 100, a parameter determination module 200 and a result determination module 300, wherein:

[0260] A model building module 100 is used to obtain a constraint parameter identification model corresponding to the equivalent circuit model based on the equivalent circuit model of the current battery, wherein the constraint parameter identification model is a parameter identification model including constraint conditions;

[0261] A parameter determination module 200, configured to obtain a first characteristic parameter of the current battery based on the constraint parameter identification model and the sequential operating condition data;

[0262] The result determination module 300 is used to obtain a parameter identification result based on the first characteristic parameter input into the neural network model.

[0263] In one embodiment, the equivalent circuit model includes: DC internal resistance, polarization internal resistance and polarization capacitance.

[0264] In one embodiment, the model building module 100 is further used to:

[0265] Based on the equivalent circuit model, constructing a transfer function corresponding to the equivalent circuit model;

[0266] Performing bilinear transformation on the transfer function to obtain a discretized equation;

[0267] Based on the discretized equation, determining a parameter identification model of the equivalent circuit model;

[0268] Boundary constraints are imposed on the parameters of the parameter identification model to obtain a constrained parameter identification model.

[0269] In one embodiment, the discretization equation includes a DC internal resistance coefficient, a polarization internal resistance coefficient, and a polarization time coefficient, and the model building module 100 is further used to:

[0270] Determining identification model parameters of the equivalent circuit model based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient; the identification model parameters include a historical voltage influence vector, a current current influence vector, a historical current influence vector and an open circuit voltage contribution vector;

[0271] Based on the identification model parameters, a parameter identification model of the equivalent circuit model is determined.

[0272] In one embodiment, the model building module 100 is further used to:

[0273] Determining a constraint parameter range of the historical voltage influence vector based on the polarization time;

[0274] Based on the DC internal resistance, determining a first constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector;

[0275] Based on the polarization internal resistance, determining a second constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector;

[0276] Based on the constraint parameter range, the first constraint relationship and the second constraint relationship, boundary constraints are performed on the parameters of the parameter identification model to obtain a constraint parameter identification model.

[0277] In one embodiment, the model building module 100 is further used to:

[0278] Determining a minimum value of the historical voltage influence vector based on the minimum value of the polarization time;

[0279] Based on the maximum value of the polarization time, a maximum value of the historical voltage influence vector is determined.

[0280] In one embodiment, determining the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector based on the DC internal resistance includes:

[0281] Based on the relationship between the DC internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the DC internal resistance, a first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

[0282] In one embodiment, the model building module 100 is further used to:

[0283] Based on the relationship between the polarization internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the polarization internal resistance, a second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

[0284] In one embodiment, the sequential operating condition data includes voltage data and current data at multiple different temperatures and charge states, and the parameter determination module 200 is further used to:

[0285] Based on the parameter identification model and the time series operating condition data, sequentially calculating the preliminary characteristic parameters of the current battery at each moment;

[0286] Performing stability verification on the preliminary characteristic parameters at each moment;

[0287] Based on the preliminary characteristic parameters at multiple moments verified by stability, the first characteristic parameters of the current battery are obtained.

[0288] In one embodiment, the result determination module 300 is further used for:

[0289] Calculating the current terminal voltage of the battery based on the first characteristic parameter;

[0290] Calculating a residual of the neural network model based on a difference between a measured voltage and the terminal voltage;

[0291] Based on the residual, optimizing the neural network model;

[0292] Based on the first characteristic parameter and the optimized neural network model, a parameter identification result is obtained.

[0293] In one embodiment, the result determination module 300 is further used for:

[0294] Inputting the first characteristic parameter into the neural network model to obtain a second characteristic parameter;

[0295] Substituting the second characteristic parameter and the monitoring operating condition data into the equivalent circuit model, the terminal voltage is obtained; the monitoring operating condition data includes historical state of charge, current and temperature.

[0296] In one embodiment, the result determination module 300 is further used for:

[0297] When the residual is greater than a preset residual threshold, the weight parameters of the neural network model are updated based on the residual, and the residual of the neural network model is recalculated;

[0298] When the residual is less than or equal to a preset residual threshold, a weight operation is performed on the first characteristic parameter based on the optimized neural network model to obtain an optimized characteristic parameter; and a parameter identification result is obtained based on the optimized characteristic parameter.

[0299] Each module in the above-mentioned battery parameter identification device can be implemented in whole or in part by software, hardware or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0300] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 7 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store timing operating condition data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a battery parameter identification method is implemented.

[0301] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0302] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the battery parameter identification method of any of the above embodiments is implemented:

[0303] Based on an equivalent circuit model of the current battery, a constraint parameter identification model corresponding to the equivalent circuit model is obtained, wherein the constraint parameter identification model is a parameter identification model including constraint conditions;

[0304] Based on the constraint parameter identification model and the time sequence operating condition data, obtaining a first characteristic parameter of the current battery;

[0305] Based on the first characteristic parameter being input into the neural network model, a parameter identification result is obtained.

[0306] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the battery parameter identification method of any of the above embodiments is implemented:

[0307] Based on an equivalent circuit model of the current battery, a constraint parameter identification model corresponding to the equivalent circuit model is obtained, wherein the constraint parameter identification model is a parameter identification model including constraint conditions;

[0308] Based on the constraint parameter identification model and the time sequence operating condition data, obtaining a first characteristic parameter of the current battery;

[0309] Based on the first characteristic parameter being input into the neural network model, a parameter identification result is obtained.

[0310] In one embodiment, a computer program product is provided. When the computer program is executed by a processor, the battery parameter identification method of any of the above embodiments is implemented:

[0311] Based on an equivalent circuit model of the current battery, a constraint parameter identification model corresponding to the equivalent circuit model is obtained, wherein the constraint parameter identification model is a parameter identification model including constraint conditions;

[0312] Based on the constraint parameter identification model and the time sequence operating condition data, obtaining a first characteristic parameter of the current battery;

[0313] Based on the first characteristic parameter being input into the neural network model, a parameter identification result is obtained.

[0314] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0315] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0316] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0317] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A battery parameter identification method, characterized in that: The battery parameter identification method comprises: Based on the equivalent circuit model of the current battery, a constrained parameter identification model corresponding to the equivalent circuit model is obtained, and the constrained parameter identification model is a parameter identification model containing constraint conditions; the equivalent circuit model includes: DC internal resistance, polarization internal resistance and polarization capacitance; the constrained parameter identification model corresponding to the equivalent circuit model based on the equivalent circuit model includes: constructing a transfer function corresponding to the equivalent circuit model based on the equivalent circuit model; performing a bilinear transformation on the transfer function to obtain a discretized equation; determining the parameter identification model of the equivalent circuit model based on the discretized equation; performing boundary constraints on the parameters of the parameter identification model to obtain a constrained parameter identification model; the discretized equation includes a DC internal resistance coefficient, a polarization internal resistance coefficient and a polarization time coefficient, and the parameter identification model of the equivalent circuit model determined based on the discretized equation includes: determining the identification model parameters of the equivalent circuit model based on the DC internal resistance coefficient, the polarization internal resistance coefficient and the polarization time coefficient; the identification model parameters include a historical voltage influence vector, a current current influence vector, a historical current influence vector and open circuit voltage contribution vector; the historical voltage influence vector is used to represent the influence of the polarization voltage at the previous moment on the current terminal voltage; the current current influence vector is used to represent the influence of the current current on the current terminal voltage, and the historical current influence vector is used to represent the influence of the current at the previous moment on the current terminal voltage; the open circuit voltage contribution vector is used to represent the contribution of the open circuit voltage to the current terminal voltage; based on the identification model parameters, the parameter identification model of the equivalent circuit model is determined; the parameters of the parameter identification model are subjected to boundary constraints to obtain the constrained parameter identification model, including: based on the polarization time, determining the constraint parameter range of the historical voltage influence vector; based on the DC internal resistance, determining the first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector; based on the polarization internal resistance, determining the second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector; based on the constraint parameter range, the first constraint relationship and the second constraint relationship, the parameters of the parameter identification model are subjected to boundary constraints to obtain the constrained parameter identification model; Based on the constraint parameter identification model and the time sequence operating condition data, obtaining a first characteristic parameter of the current battery; Based on the first characteristic parameter being input into the neural network model, a parameter identification result is obtained.

2. The battery parameter identification method according to claim 1, characterized in that: The step of determining the constraint parameter range of the historical voltage influence vector based on the polarization time includes: Determining a minimum value of the historical voltage influence vector based on the minimum value of the polarization time; Based on the maximum value of the polarization time, a maximum value of the historical voltage influence vector is determined.

3. The battery parameter identification method according to claim 1, characterized in that: The determining, based on the DC internal resistance, a first constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector comprises: Based on the relationship between the DC internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the DC internal resistance, a first constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

4. The battery parameter identification method according to claim 1, characterized in that: The determining, based on the polarization internal resistance, a second constraint relationship between the historical voltage influence vector, the current current influence vector, and the historical voltage influence vector comprises: Based on the relationship between the polarization internal resistance and the historical voltage influence vector, the relationship between the current current influence vector and the historical voltage influence vector, and the minimum value of the polarization internal resistance, a second constraint relationship between the historical voltage influence vector, the current current influence vector and the historical voltage influence vector is determined.

5. The battery parameter identification method according to claim 1, characterized in that: The sequential operating condition data includes voltage data and current data at multiple different temperatures and charge states. The first characteristic parameter of the current battery obtained based on the parameter identification model and the sequential operating condition data includes: Based on the parameter identification model and the time series operating condition data, sequentially calculating the preliminary characteristic parameters of the current battery at each moment; Performing stability verification on the preliminary characteristic parameters at each moment; Based on the preliminary characteristic parameters at multiple moments verified by stability, the first characteristic parameters of the current battery are obtained.

6. The battery parameter identification method according to claim 1, characterized in that: The step of inputting the first characteristic parameter into the neural network model to obtain a parameter identification result includes: Calculating the current terminal voltage of the battery based on the first characteristic parameter; Calculating a residual of the neural network model based on a difference between a measured voltage and the terminal voltage; Based on the residual, optimizing the neural network model; Based on the first characteristic parameter and the optimized neural network model, a parameter identification result is obtained.

7. The battery parameter identification method according to claim 6, characterized in that: The calculating the current terminal voltage of the battery based on the first characteristic parameter includes: Inputting the first characteristic parameter into the neural network model to obtain a second characteristic parameter; Substituting the second characteristic parameter and the monitoring operating condition data into the equivalent circuit model, the terminal voltage is obtained; the monitoring operating condition data includes historical state of charge, current and temperature.

8. The battery parameter identification method according to claim 6, characterized in that: Based on the residual, optimizing the neural network model; Based on the first characteristic parameter and the optimized neural network model, the parameter identification result is obtained including: When the residual is greater than a preset residual threshold, the weight parameters of the neural network model are updated based on the residual, and the residual of the neural network model is recalculated; When the residual is less than or equal to a preset residual threshold, a weight operation is performed on the first characteristic parameter based on the optimized neural network model to obtain an optimized characteristic parameter; and a parameter identification result is obtained based on the optimized characteristic parameter.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

11. A computer program product, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images