Search

CN-121978603-A - Sectional type online SOC dynamic calibration method with forgetting factor, computer equipment and electric automobile

CN121978603ACN 121978603 ACN121978603 ACN 121978603ACN-121978603-A

Abstract

The invention relates to the technical field of battery management and discloses a sectional type on-line SOC dynamic calibration method with forgetting factors, computer equipment and an electric automobile; according to the on-line SOC dynamic calibration method, the battery parameter table is constructed, the least square method with the forgetting factors is adopted for on-line parameter identification, the forgetting factors are dynamically adjusted through the segmentation error threshold, the numerical stability is ensured by combining overrun judgment, the SOC is estimated and calibrated in real time by using the extended Kalman filtering algorithm, the accumulated error is effectively restrained, the estimation precision and the system adaptability are remarkably improved, and the method is more suitable for the actual running conditions of automobiles.

Inventors

  • ZHAO ZHIPENG
  • YANG YANHUI
  • JI XIANG
  • ZENG GUOJIAN
  • CAI HUAJUAN

Assignees

  • 安徽锐能科技有限公司

Dates

Publication Date
20260505
Application Date
20251231

Claims (9)

  1. 1. The segmented online SOC dynamic calibration method with the forgetting factor is characterized by comprising the following steps of: Constructing a battery model parameter table; identifying battery model parameters on line according to a least square method with forgetting factors; and (5) estimating and calibrating the SOC on line by adopting an EKF method.
  2. 2. The dynamic calibration method according to claim 1, wherein constructing a battery model parameter table comprises: acquiring current, voltage and time series data of the battery under different working conditions by adopting an HPPC experimental method, establishing an equivalent circuit model according to a formula group (1), ,(1) Wherein, the Is that The battery terminal voltage at the moment in time, Is that The open circuit voltage at the moment in time, Is that The first polarization voltage at the moment in time, Is that The second polarization voltage at the moment in time, Is that The first polarization voltage at the moment in time, Is that The second polarization voltage at the moment in time, Is the ohmic internal resistance, For the first polarization resistance, For the second polarization resistance, the first polarization resistance, For the first polarized capacitance of the capacitor, For the second polarized capacitance of the capacitor, Is that The current at the moment in time is, Is that The current at the moment in time is, Is a natural constant which is used for the production of the high-temperature-resistant ceramic material, Is the time step; determining an SOC breakpoint according to the change trend of the current and the time; Dividing data into a plurality of data subsets according to breakpoints, and determining a plurality of SOC intervals; performing battery response curve fitting on each data subset, and calculating to obtain ohmic internal resistance parameters and polarization parameters; And performing n-order polynomial fitting on the open circuit voltage and the SOC to generate a battery model parameter table.
  3. 3. The dynamic calibration method of claim 2, wherein performing a battery response curve fit for each subset of data, calculating the obtained ohmic internal resistance parameter and the polarization parameter, comprises: identifying and extracting key curve segments of a battery response curve; Fitting a curve segment of an instantaneous voltage jump point in the impulse response by adopting a formula (2) to obtain ohmic resistance parameters, ,(2) Wherein, the 、 、 、 As a voltage point in the HPPC experiment, Is a pulse current; fitting the relaxation curve segment by adopting a formula (3) to obtain polarization parameters, ,(3) Wherein, the For the battery terminal voltage to be the same, Is the open circuit voltage of the power supply, In the event of a current flow, Is a natural constant which is used for the production of the high-temperature-resistant ceramic material, For the time step size of the time step, For the first polarization resistance, For the second polarization resistance, the first polarization resistance, For the first polarized capacitance of the capacitor, Is the second polarized capacitance.
  4. 4. The dynamic calibration method according to claim 1, wherein identifying battery model parameters on-line according to a least squares method with forgetting factors comprises: Constructing a continuous S-domain transfer function based on an equivalent circuit model by adopting a formula (4), ,(4) Wherein, the As a transfer function of the S-domain, As a first time constant, the first time constant, For the second time constant to be a second time constant, Is a Laplace variable; a discretized transfer function is constructed according to bilinear variations using equation (5), ,(5) Wherein, the Is a constant value, and is used for the treatment of the skin, Is a Z-transformed variable; respectively defining a parameter vector to be estimated and an input data amount by adopting formulas (6) - (7), ,(6) ,(7) - ,(8) Wherein, the Is that The parameter vector to be estimated for the moment in time, 、 、 、 、 、 Is a constant value, and is used for the treatment of the skin, Is that The amount of input data at the moment in time, Is that The output variable of the moment of time, Is that The output variable of the moment of time, Is that The output variable of the moment of time, Is that The input variable of the moment of time, Is that The input variable of the moment of time, Is that The input variable of the moment of time, Is that The output variable of the time is The difference between the open circuit voltage and the terminal voltage at the moment; Linearizing according to a least square method and obtaining an output variable according to a formula (9), ,(9) Wherein, the Is that The output variable of the moment of time, Is that The transpose of the input data vector at time instant, Is that The parameter vector to be estimated for the moment in time, Is that Prediction error of time; updating the estimated parameter vector according to the formula set (10) by adopting a recursive algorithm, ,(10) Wherein, the Is that The prediction error of the moment in time, Is that The transpose of the input data vector at time instant, Is that The input data vector of the moment in time, Is that The estimated parameter vector of the moment in time, Is that The estimated parameter vector of the moment in time, Is that The kalman gain matrix of the moment in time, Is that The covariance matrix of the time of day, Is that The covariance matrix of the time of day, As a parameter of the forgetting factor, Is a unit matrix; different allowable errors are set according to different SOC intervals, a formula (11) is adopted to dynamically adjust forgetting factor parameters in a sectional mode, ,(11) Wherein, the Is that The forgetting factor parameter of the moment of time, As a minimum forgetting factor parameter, As a parameter of the maximum forgetting factor, In order for the coefficient of variation to be a function of, Is that The error scaling factor for the time of day, Is an allowable error threshold; identifying a constant in the discretization transfer function according to the dynamic forgetting factor parameters, acquiring an intermediate variable by adopting a formula group (12), and processing an overrun value in the intermediate variable by adopting a parameter overrun processing method; ,(12) Wherein, the 、 、 、 、 As an intermediate variable, T is sampling time; performing parameter conversion by adopting a formula (13) according to the intermediate variable to obtain battery model parameters, ,(13) Wherein, the As a first time constant, the first time constant, Is a second time constant.
  5. 5. The method of dynamic calibration according to claim 4, wherein the processing of overrun values in intermediate variables using a parameter overrun processing method comprises: judging whether the value of the denominator item of the intermediate variable is smaller than a preset first threshold value or not; Acquiring the value of the denominator item of the intermediate variable at the previous moment under the condition that the value of the denominator item of the intermediate variable is smaller than a preset first threshold value; And under the condition that the value of the denominator of the intermediate variable is larger than or equal to a preset first threshold value, acquiring the value of the denominator of the intermediate variable at the current moment.
  6. 6. The dynamic calibration method according to claim 1, wherein the on-line estimation and calibration of the SOC using the EKF method comprises: Judging whether an EKF calibration SOC condition is triggered or not; under the condition of triggering an EKF calibration SOC condition, estimating and calibrating the SOC on line; Judging whether an EKF termination calibration SOC condition is triggered; Under the condition that the EKF is not triggered to terminate the SOC calibration condition, the steps of online estimation and SOC calibration are iterated and carried out again; and under the condition that the EKF is triggered to terminate the calibration SOC condition, outputting the current calibrated SOC.
  7. 7. The dynamic calibration method according to claim 6, wherein in case of triggering an EKF calibration SOC condition, estimating and calibrating the SOC online comprises: setting initial state and noise characteristics of the EKF algorithm, and initializing state vector according to formula (14), ,(14) Wherein, the As a state vector of the state vector, In order to achieve the initial state of charge, For the first polarization voltage to be applied, For the second polarization voltage to be applied, For the state estimation covariance matrix, In order to process the noise covariance matrix, Measuring a noise covariance matrix; predicting a state vector and an error covariance matrix of the battery model at the current time using equations (15) - (16), ,(15) ,(16) Wherein, the Is that A priori estimates of the state vector at time instants, Is that A state transition matrix of the moment of time, Is that The updated optimal estimate of the state vector at time, Is that An input matrix of the time of day, Is that The current is measured at the moment in time, Is that The predicted value of the error covariance matrix at the moment, Is that The transpose of the state transition matrix at time, A process noise covariance matrix; correcting the predicted state vector by adopting a formula (17); ,(17) Wherein, the Is that The predicted value of the terminal voltage at the moment, Is that The output matrix of the time of day, Is a feedforward matrix; the innovation vector and gain matrix are obtained using a set of formulas (18), ,(18) Wherein, the In order to create the vector of the innovation, Is that The input cell voltage value at the moment in time, Is that A gain matrix of the time of day, Is that The transposed matrix of the output matrix of time instants, Measuring a noise covariance matrix; Updating and calibrating the state vector and the error covariance matrix by adopting formulas (19) - (20) according to the innovation vector and the gain matrix to obtain the optimal estimated value at the current moment, ,(19) ,(20) Wherein, the Is that The updated optimal estimate of the state vector at time, For after updating Error covariance matrix of time.
  8. 8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the dynamic calibration method of any one of claims 1 to 7 when executing the computer program.
  9. 9. An electric vehicle, characterized in that it comprises a battery system for performing the dynamic calibration method according to any one of claims 1 to 7.

Description

Sectional type online SOC dynamic calibration method with forgetting factor, computer equipment and electric automobile Technical Field The invention relates to the technical field of battery management, in particular to a sectional type on-line SOC dynamic calibration method with forgetting factors, computer equipment and an electric automobile. Background In recent years, electric automobiles and related technologies have been rapidly developed. The estimation of the state of charge of the battery is an important component of the battery management system, and accurate SOC estimation is beneficial to fully playing the power performance of the battery system, preventing the overcharge and overdischarge of the power battery, and guaranteeing the service life of the power battery and the safety in the use process. The current common SOC estimation method comprises an ampere-hour integration method, an open circuit voltage method, a neural network method and a Kalman filtering method. The open circuit voltage method takes the terminal voltage of an electric automobile after long-time standing as open circuit voltage, determines a calibration value through the corresponding relation between the open circuit voltage and the SOC, is difficult to realize dynamic estimation in the actual operation process, a battery is used as a complex nonlinear system to estimate the SOC by adopting a neural network method, higher precision can be realized, but the SOC is difficult to realize on a singlechip due to complex calculation and a large amount of data storage space, the Kalman filtering method is used for estimating the SOC on the basis of a battery equivalent circuit model, the SOC estimation value is calculated through a system state equation, and then the estimation value is corrected according to the current voltage measurement value, so that the process of estimating the minimum variance of the system state is realized. Disclosure of Invention In order to overcome the technical problems, the invention provides a sectional type on-line SOC dynamic calibration method with forgetting factors, which is characterized in that a battery parameter table is constructed, a least square method with the forgetting factors is adopted to conduct on-line parameter identification, the forgetting factors are dynamically adjusted through a sectional error threshold, the numerical stability is ensured by combining overrun judgment, the SOC is estimated and calibrated in real time by utilizing an extended Kalman filtering algorithm, the accumulated error is effectively restrained, the estimation precision and the system adaptability are obviously improved, and the method is more suitable for the actual running working conditions of automobiles. The first aspect of the invention provides a segmented online SOC dynamic calibration method with forgetting factors, which comprises the following steps: Constructing a battery model parameter table; identifying battery model parameters on line according to a least square method with forgetting factors; and (5) estimating and calibrating the SOC on line by adopting an EKF method. Preferably, constructing the battery model parameter table includes: acquiring current, voltage and time series data of the battery under different working conditions by adopting an HPPC experimental method, establishing an equivalent circuit model according to a formula group (1), ,(1) Wherein, the Is thatThe battery terminal voltage at the moment in time,Is thatThe open circuit voltage at the moment in time,Is thatThe first polarization voltage at the moment in time,Is thatThe second polarization voltage at the moment in time,Is thatThe first polarization voltage at the moment in time,Is thatThe second polarization voltage at the moment in time,Is the ohmic internal resistance,For the first polarization resistance,For the second polarization resistance, the first polarization resistance,For the first polarized capacitance of the capacitor,For the second polarized capacitance of the capacitor,Is thatThe current at the moment in time is,Is thatThe current at the moment in time is,Is a natural constant which is used for the production of the high-temperature-resistant ceramic material,Is the time step; determining an SOC breakpoint according to the change trend of the current and the time; Dividing data into a plurality of data subsets according to breakpoints, and determining a plurality of SOC intervals; performing battery response curve fitting on each data subset, and calculating to obtain ohmic internal resistance parameters and polarization parameters; And performing n-order polynomial fitting on the open circuit voltage and the SOC to generate a battery model parameter table. Preferably, performing battery response curve fitting on each data subset, and calculating to obtain an ohmic internal resistance parameter and a polarization parameter, including: identifying and extracting key curve segments of a battery