CN-122017587-A - Battery hybrid modeling method

CN122017587ACN 122017587 ACN122017587 ACN 122017587ACN-122017587-A

Abstract

The invention relates to a battery hybrid modeling method which comprises two parts of current preprocessing and battery hybrid model modeling, wherein firstly, abnormal value preprocessing is carried out on collected battery current signals, correction and reconstruction are carried out on current data in combination with a vehicle running state so as to obtain normal current input capable of representing a real working state of a battery, and then, a normal voltage prediction model formed by fusing a physical mechanism model, a circuit equivalent model and a data driving model is constructed and used for accurately predicting battery terminal voltage under different running working conditions.

Inventors

LI DA
LI WEILIN
LIU PEI
LIU SONGYAN
WU YU
CHENG BOYUAN
WANG NINGHAO
REN YIFENG

Assignees

西北工业大学

Dates

Publication Date: 20260512
Application Date: 20260206

Claims (5)

1. A battery hybrid modeling method comprises current pretreatment and battery hybrid model modeling, and is characterized in that, The current preprocessing comprises the steps of detecting an abnormal value of a battery current signal, and replacing or reconstructing the abnormal current according to the running state of the battery when the abnormal current is detected; The running state at least comprises a vehicle running state and a charging state, and different current correction modes are adopted in different running states; The mixed model modeling comprises the steps of respectively constructing a normal voltage prediction model formed by fusing a physical mechanism model, a circuit equivalent model and a data driving model based on normal current input obtained by current pretreatment, and accurately predicting the voltage of a battery terminal under different operation conditions.
2. The battery hybrid modeling method according to claim 1, wherein the battery current is preprocessed based on a motor bus current and a motor efficiency equation in a running state: calculating and judging whether the total current error value is in a normal interval, wherein the total current error value is calculated according to the following formula: In the formula, The battery current is the battery current at the time t, The actual motor bus current at time t, Accessory current at time t; When the total current error is in the normal interval, the battery current at the time t is taken as the battery current after the processing at the time t: When the total current error is not in the normal interval, continuously judging whether the difference value between the measured value of the motor bus current and the estimated value of the motor bus current is in the normal interval, wherein the difference value between the estimated value of the motor bus current and the actual motor bus current is expressed as: In the formula, The estimated value of the bus current of the motor at the moment t is calculated by a motor efficiency equation, The actual bus current of the motor at the moment t; The equation for motor efficiency is: In the formula, Is the voltage of the motor and is used for controlling the motor, Is the output torque of the motor and is used for controlling the motor, Is the output rotating speed of the motor, Actual motor bus current; the calculated estimated value of the bus current of the motor at the time t is as follows: In the formula, The torque output by the motor at the moment t, The output rotating speed of the motor at the moment t, The motor voltage at the moment t; difference between motor bus current estimate and measurement In the normal interval, taking the sum of the actual motor bus current and the accessory current as the processed current: otherwise, taking the sum of the motor bus current estimated value and the accessory current as the processed current: 。
3. the battery hybrid modeling method of claim 1, wherein the battery current is preprocessed in a charged state based on a nominal charging current: The total current error value is calculated according to the following equation: In the formula, The charging current is calibrated for the time t, The battery current is the t moment; For the case where the total current error value is in the normal interval, the battery current is taken as the processed current: For the case that the total current error value is not in the normal interval, the calibration charging current is taken as the processed current: 。
4. The battery hybrid modeling method according to claim 1, wherein the abnormal current is replaced by the normal current after the current collected by the real vehicle sensor is processed by the current preprocessing model, so that the hybrid model predicts the accurate normal output voltage.
5. The battery hybrid modeling method according to claim 1, wherein the hybrid model modeling specifically includes: Based on a single particle model, the battery is input with current And output voltage The relationship between is described as: In the formula, And Respectively positive and negative electrode balance potentials, And Respectively the overpotential of the positive electrode and the negative electrode, And The lithium ion concentration on the surface of the positive and negative solid phase particles is respectively, And The reaction kinetics rates of the positive electrode and the negative electrode are respectively, And The molar flux of the surfaces of positive and negative particles respectively, And Respectively positive and negative electrode solid phase electrolyte membrane resistance, And The specific interface areas of the positive electrode and the negative electrode are respectively, And The thicknesses of the positive electrode and the negative electrode are respectively; the integration of the above-mentioned components of the formula of the relationship between the battery input current and output voltage can be achieved: Wherein U is the terminal voltage, Is the balance potential of the positive electrode and the negative electrode, Is the difference between the overpotential of the positive electrode and the negative electrode, Is the potential difference of the solid electrolyte interphase membrane; calculating the potential difference of the solid electrolyte interphase membrane by adopting an electrochemical model Simultaneously adopting an electrochemical model to calculate the surface molar flux of positive and negative particles And As a result of Is input into: In the formula, For the current after CPM processing by the current processing module, F is the faraday constant, a is the surface area of the electrode, And The specific interface areas of the positive electrode and the negative electrode are respectively, And The thicknesses of the positive electrode and the negative electrode are respectively; Estimation using the Thevenin model Represented by the following formula: In the formula, Open circuit voltage for the battery; According to kirchhoff's law, the circuit equation of the davin model is as follows: In the formula, 、、、 I is respectively polarization voltage, internal resistance, polarization capacitance and current; radial basis function neural network fitting is adopted to calculate ; The input positive and negative particle surface mole flux And Mapping to high-dimensional space by radial basis functions, followed by linear output layer computation ; The transfer function between the input layer and the RBF hidden layer is: where exp () is an exponential function, Width parameters that are radial basis functions; the calculation formula of the output layer is as follows: where m is the number of hidden layer neurons, Is the weight of the hidden layer and the output layer; based on the mean square error, defining a loss function of the RBF hidden layer: In the formula, Is the true value of the difference between the positive and negative overpotential at the time t, A predicted value of the difference between the positive and negative overpotential at the time t; and synchronously optimizing parameters in the loss function of the RBF hidden layer by adopting a gradient descent method: where α is the learning rate.

Description

Battery hybrid modeling method Technical Field The invention belongs to the field of electrical engineering, and particularly relates to a battery hybrid modeling method. Background The power battery is widely applied to various electric equipment such as electric automobiles, aviation battery systems and the like, and the battery terminal voltage is influenced by various factors such as input current, working environment, internal electrochemical process and the like in the operation process, so that obvious nonlinearity and time-varying characteristics are presented. The existing battery modeling method mainly comprises an electrochemical model, an equivalent circuit model and a data driving model, different models have advantages and limitations in the aspects of physical mechanism description capacity, calculation complexity, nonlinear modeling precision and the like, and a single model is difficult to simultaneously consider modeling precision and calculation efficiency under complex working conditions. In addition, in the actual operation process, the battery current measurement is easy to be influenced by sensor noise, signal interference or abnormal data, and if abnormal current is directly input as a model, the voltage prediction result is easy to be unstable and even larger deviation is generated. Therefore, a high-precision modeling method for the normal terminal voltage of the battery, which is suitable for various application scenes by integrating the advantages of various models on the basis of ensuring the effectiveness of current input, is needed. Disclosure of Invention The invention provides a battery hybrid modeling method which comprises the steps of preprocessing a battery current signal, correcting and reconstructing current data by combining a vehicle running state to obtain normal current input capable of reflecting a real working state of a battery, and constructing a normal voltage prediction model formed by fusing a physical mechanism model, a circuit equivalent model and a data driving model on the basis of the current preprocessing, the battery current signal preprocessing and the correction and the reconstruction of the current data, wherein the normal voltage prediction model is used for accurately predicting the voltage of a battery terminal under different running working conditions. The battery current preprocessing is used for constructing a current abnormal value preprocessing model aiming at the problems of abnormal current collected by a vehicle-mounted sensor and the like caused by sensor faults and data transmission errors, and ensuring that the input current of the coupling model is a normal value. The vehicle state is divided into two states of running and charging, and different current processing flows are adopted under different running states. In a driving state, preprocessing the battery current based on a motor bus current and a motor efficiency equation: calculating and judging whether the total current error value is in a normal interval, wherein the total current error value is calculated according to the following formula: In the formula, The battery current is the battery current at the time t,The actual motor bus current at time t,The accessory current at time t. When the total current error is in the normal interval, the battery current at the time t is taken as the battery current after the processing at the time t: When the total current error is not in the normal interval, continuously judging whether the difference value between the measured value of the motor bus current and the estimated value of the motor bus current is in the normal interval, wherein the difference value between the estimated value of the motor bus current and the actual motor bus current is expressed as: In the formula, The estimated value of the bus current of the motor at the moment t is calculated by a motor efficiency equation,The actual motor bus current at time t. The equation for motor efficiency is: In the formula, Is the voltage of the motor and is used for controlling the motor,Is the output torque of the motor and is used for controlling the motor,Is the output rotating speed of the motor,Is the actual motor bus current. The calculated estimated value of the bus current of the motor at the time t is as follows: In the formula, The torque output by the motor at the moment t,The output rotating speed of the motor at the moment t,Is the motor voltage at time t. Difference between motor bus current estimate and measurementIn the normal interval, taking the sum of the actual motor bus current and the accessory current as the processed current: otherwise, taking the sum of the motor bus current estimated value and the accessory current as the processed current: in a charged state, the battery current is preprocessed based on a nominal charging current. The total current error value is calculated according to the following equation: In the formula, The charging current is c