CN-122017656-A - Battery module state estimation method considering dynamic characteristics of functional state

CN122017656ACN 122017656 ACN122017656 ACN 122017656ACN-122017656-A

Abstract

The invention provides a battery module state estimation method considering dynamic characteristics of functional states, and relates to the technical field of battery management. The method comprises the steps of collecting characteristic parameters of a single battery, calculating polarization voltage in the battery, constructing a charge state model representing short-time scale characteristics of the battery and a health state model representing long-time scale characteristics of the battery according to factors affecting state estimation of a battery module, further establishing a functional state dynamic model of the single battery and a state estimation model under double time scales, expanding the functional state dynamic model of the single battery and the state estimation model under the double time scales to a battery module level, and establishing the functional state dynamic model of the battery module and the state estimation model under the double time scales, so that state estimation values and the functional states of the battery module are calculated. The invention effectively improves the accuracy and dynamic response capability of battery state estimation and provides a basis for long-term health management of batteries.

Inventors

YAN NING
HE ZILUN
ZHAO LU
GAO LEI

Assignees

沈阳工业大学

Dates

Publication Date: 20260512
Application Date: 20260304

Claims (10)

1. The battery module state estimation method considering the dynamic characteristics of the functional state is characterized by comprising the following steps: collecting characteristic parameters of the single battery, and calculating electrochemical polarization voltage and concentration polarization voltage inside the battery; respectively constructing a state-of-charge model representing short-time scale characteristics of the battery and a state-of-health model representing long-time scale characteristics of the battery according to factors influencing state estimation of the battery module; based on the electrochemical polarization voltage, the concentration polarization voltage, the state of charge model and the state of health model, a functional state dynamic model of the single battery is established; Based on the state-of-charge model and the state-of-health model, constructing a state estimation model of the single battery under a double time scale; According to the functional state dynamic model of the single battery, constructing a functional state dynamic model of the battery module; according to a state estimation model of the single battery under the double time scales and a functional state dynamic model of the battery module, constructing a state estimation model of the battery module under the double time scales; According to the state estimation model of the battery module under the double time scales, the state of charge estimation value and the state of health estimation value of the battery module at the current moment are calculated, and then according to the functional state dynamic model of the battery module, the functional state of the battery module at the current moment is calculated.
2. The method for estimating a battery module according to claim 1, wherein said characteristic parameters of said battery cell include voltage, current, internal resistance and capacitance.
3. The method for estimating the state of the battery module according to claim 1, wherein the specific method for respectively constructing the state-of-charge model representing the short-time scale characteristic of the battery and the state-of-health model representing the long-time scale characteristic of the battery according to the factors affecting the state estimation of the battery module is as follows: Determining factors influencing the state estimation of the battery module, including voltage, current, internal resistance, capacitance, state of charge and health state; Based on an ampere-hour integration method, establishing a state-of-charge dynamic model for representing the short-time scale characteristics of the battery; ; Wherein S OC (t) is the current time Is a state of charge of (2); For the current moment Is set in the above-mentioned state; Is the rated capacity of the single battery; is coulombic efficiency; For the current moment Is a health state of (a); establishing a health state evolution model for representing long-time scale characteristics of a battery, wherein the health state evolution model comprises a health state evolution model considering battery capacity attenuation and a health state evolution model considering battery internal resistance increase; the state of health evolution model considering battery capacity decay is: ; wherein S OH,Q is a capacity-based health status; the number of complete cycles experienced by the cell; the average working temperature of the single battery; is the capacity fade coefficient; empirical parameters for capacity fade as a function of cycle number; is the activation energy of the aging reaction; is a universal gas constant; The state of health evolution model considering the increase of the internal resistance of the battery is as follows: ; Wherein, the Ohmic internal resistance in the current aging state; Initial ohmic internal resistance of the single battery; is an internal resistance growth coefficient; is an empirical parameter of internal resistance increase with cycle number; is the activation energy in the internal resistance increasing process; Is a health state based on internal resistance in the current aging state.
4. The method for estimating the state of a battery module according to claim 3, wherein the specific method for establishing the dynamic model of the functional state of the single battery based on the electrochemical polarization voltage, the concentration polarization voltage, the state of charge model and the state of health model is as follows: Based on the electrochemical polarization voltage and the concentration polarization voltage, establishing a dynamic response equation of terminal voltage according to the polarization effect inside the battery; Based on the dynamic response equation of the terminal voltage, establishing a prediction model of the terminal voltage by iteratively calculating the terminal voltage at the current moment; Establishing a maximum current constraint model based on the voltage limit and the temperature limit of the battery; Limiting power of the battery under two working conditions of discharging and charging is respectively defined; Based on the prediction model of the terminal voltage, the maximum current constraint model and the defined limit power, a functional state dynamic model of the single battery is established, and the functional state dynamic model is expressed as follows: ; Wherein, the Maximum discharge power for the functional state; maximum charging power for the functional state; For duration of time; a minimum allowable voltage for battery operation; the highest allowable voltage for battery operation; Is the maximum allowable discharge current; Is the maximum charge current allowed.
5. The method for estimating the state of the battery module according to claim 4, wherein the specific method for constructing the state estimation model of the single battery under the double time scales based on the state of charge model and the state of health model is as follows: Respectively defining state vectors of the single battery under a short time scale and a long time scale based on the state-of-charge model and the state-of-health model; according to the definition of the state vector of the single battery in a short time scale and a long time scale, a state transfer function of a time scale is established; And according to the state transfer function of the time scale, carrying out recursive estimation on the state vector of the single battery under the short time scale and the state vector under the long time scale by using an observation residual error correction mechanism to obtain the optimal estimated value of the state vector of the single battery under the short time scale and the long time scale, and completing dynamic modeling of the functional state of the single battery.
6. The method for estimating the state of the battery module according to claim 5, wherein the method for recursively estimating the state vector of the single battery in a short time scale and the state vector in a long time scale by using the observation residual correction mechanism according to the state transfer function in the time scale is as follows: according to the state transfer function of the time scale, a state estimation model of the single battery in a short time scale and a long time scale is established; for the short time scale, the sampling moment is sampled based on the last short time scale State vector optimal estimation value and current long time scale sampling moment of (2) According to the state vector estimation value of the single battery short time scale, obtaining the current short time scale sampling moment State vector predictors of (a); Obtaining the current short time scale sampling time of single battery Calculates the observation residual error, and then samples the current short time scale according to the state update equation of the single battery short time scale Updating the state vector predicted value of (2) to obtain the current short time scale sampling moment Is a state vector optimal estimation value; for long time scale, based on last long time scale sampling time State vector optimal estimation value and current long time scale sampling moment of (2) According to the state prediction equation of the long time scale of the single battery, the current sampling moment of the long time scale is obtained State vector predictors of (a); using sampling instants from the last long time scale Up to the current long time scale sampling instant The internally accumulated observation residual vectors are used for constructing virtual observation increments of a long time scale; according to the state updating equation of the long time scale of the single battery, the current sampling moment of the long time scale is sampled by utilizing the virtual observation increment of the long time scale Updating the state vector predicted value of (2) to obtain the current long-time scale sampling moment Is a state vector optimal estimate of (1).
7. The method for estimating a state of a battery module according to claim 6, wherein the specific method for constructing the dynamic model of the functional state of the battery module according to the dynamic model of the functional state of the battery cell comprises: For the first in the battery module The single battery is used for respectively calculating the deviation rate of the rated capacity and the deviation rate of the ohmic internal resistance of the single battery; calculating the voltage safety margin deviation of the ith single battery according to the deviation rate of the rated capacity and the deviation rate of the ohmic internal resistance based on the module inconsistency parameter matrix; based on the voltage safety margin deviation, combining the voltage distribution of the single batteries to construct a battery module terminal voltage constraint model; Definition of the first embodiment An inconsistency factor of the individual cells; And constructing a functional state dynamic model of the battery module based on the battery module terminal voltage constraint model and the inconsistency factors of all the single batteries.
8. The method for estimating the state of the battery module according to claim 7, wherein the specific method for constructing the state estimation model of the battery module in the double time scales according to the state estimation model of the single battery in the double time scales and the functional state dynamic model of the battery module comprises the following steps: Constructing a state vector of the battery module according to a state estimation model of the single battery under the double time scales; Defining a system topology matrix of the battery module according to the electric connection mode of the battery module; Establishing a module-level constraint equation based on a system topology matrix of the battery module; ; Wherein, the The current vector is the current vector of the single battery; a system current topology matrix; is the total current of the module; is a current constraint error term; Is the terminal voltage of the battery module; is a system voltage topology matrix; Is the voltage vector of the single battery; Is a voltage constraint error term; performing recursive estimation on the state vector of the battery module in a short time scale and the state vector of the battery module in a long time scale according to the module-level constraint equation to obtain optimal estimated values of the state vector of the battery module in the short time scale and the long time scale; acquiring according to the optimal estimated value of the state vector of the battery module in a short time scale The state of charge estimate of each cell at the moment; acquiring according to the optimal estimated value of the state vector of the battery module under the long time scale State of health estimation for each cell at the moment.
9. The method for estimating the state of the battery module according to claim 8, wherein the specific method for recursively estimating the state vector of the battery module in a short time scale and the state vector in a long time scale according to the module-level constraint equation to obtain the optimal estimated value of the state vector of the battery module in the short time scale and the long time scale comprises the following steps: based on the state prediction equation of the battery cell in a short time scale, constructing the state prediction equation of the battery module in a short time scale to obtain the sampling moment of the battery module in the current short time scale State vector predictors of (a); ; Wherein, the Is the first Sampling time of each single battery in current short time scale State vector predictors of (a); is the first A short time scale state transfer function of the individual cells; is the first The single battery is arranged at A state vector optimal estimated value at the moment; sampling time of the ith single battery in the current long time scale State vector predictors of (a); a topology interaction gain matrix; Is a topology state interaction item; Is the first to A set of adjacent single cells; Is the first to Adjacent ones of the single cells A plurality of single batteries; Is a topological weight coefficient; for the j-th adjacent single battery A state vector estimate of time; acquiring current short time scale sampling time of battery module And calculates the observation residual vector of the battery module; According to a state updating equation of the battery module in a short time scale, using an observation residual vector of the battery module to sample the battery module at the current short time scale Updating the state vector predicted value of (2) to obtain the sampling time of the battery module at the current short time scale Is a state vector optimal estimation value; ; Wherein, the Is the first The sampling time of each single battery in the current short time scale Is a state vector optimal estimation value; is the first A Kalman gain matrix of each adjacent single battery at the time t; is the first The observation residual vectors of the single batteries are; a coordination matrix for topology observation; Is the first to Observing residual vectors by a set of adjacent single batteries; the intensity coefficient is topologically coordinated; is the first A topological weight matrix of the individual cells; Based on a state prediction equation of a single battery in a long time scale, constructing a global state update equation of the battery module to obtain the sampling moment of the battery module in the current long time scale Is a state vector optimal estimation value; ; Wherein, the For the sampling time of the battery module at the current long time scale State vector estimates of (2); is the first Current long time scale sampling time of single battery Is a state vector optimal estimation value; is a fusion gain matrix; Is a consistency correction amount; An arithmetic mean vector of state vector estimates for each cell long time scale; Is a battery module A state vector estimate of time; is the number of series/parallel single batteries in the battery module.
10. The method for estimating the state of a battery module according to claim 9, wherein the specific method for calculating the state of charge and the state of health at the current time of the battery module according to the state estimation model of the battery module in the dual time scale, and further calculating the state of function at the current time of the battery module according to the state dynamic model of the battery module is as follows: Defining a weighted average value of the states of charge of all the single batteries in the battery module, and calculating a state of charge estimated value and a state of health estimated value of the battery module at the current moment according to a state estimation model of the battery module under a double time scale; ; Wherein, the For the sampling time of the battery module at the current long time scale Is a health state estimate of (1); And Are all weight coefficients; is the first The single battery is arranged at State of health based on capacity at time; is the first The single battery is arranged at A state of health based on the internal resistance at the moment; For the sampling time of the battery module in the current short time scale Is a state of charge estimate for (1); is the first The weight coefficient of the short time scale of each single battery; is the first The single battery is arranged at A state of charge estimate at time; Calculating the maximum charging power and the maximum discharging power of the battery module at the current moment according to the dynamic model of the functional state of the battery module based on the state of charge estimated value and the state of health estimated value of the battery module at the current moment; ; Wherein, the Maximum sustainable discharge power for the functional state of the battery module; Maximum sustainable charging power for the functional state of the battery module; Is the minimum allowable voltage of the single battery; is the maximum allowable voltage of the single battery; For duration of time; To at the same time Time, the first The single battery is arranged at Maximum allowable discharge current; To at the same time Time, the first The single battery is arranged at Maximum allowable charging current; is the first And the inconsistency factor of the single batteries.

Description

Battery module state estimation method considering dynamic characteristics of functional state Technical Field The invention relates to the technical field of battery management, in particular to a battery module state estimation method considering dynamic characteristics of functional states. Background Along with the acceleration of the global new energy automobile industrial scale process, the precision of the state estimation of the power battery system has become a core technical bottleneck for restricting the improvement of the energy efficiency, the safety and the service life of the whole automobile. The high-precision state estimation is important to optimize the battery energy management and prevent overcharge and overdischarge, and is also an important basis for realizing early warning of faults and prolonging the service life of the battery. Researches show that the accurate state estimation can improve the prediction accuracy of the continuous voyage mileage of the electric automobile by about 15% -20%, and the failure occurrence rate of a battery system is remarkably reduced. Under this background, the intelligent development of the battery management system puts higher demands on the accuracy and dynamic response speed of the state estimation technology in the whole life cycle. However, the existing battery module state estimation method still faces significant challenges. The traditional method focuses on single batteries, and dynamic characteristics of battery functional states along with temperature, aging, multiplying power and other conditions cannot be fully considered, so that the peak power prediction error is generally more than 10%. At the module level, the existing estimation strategy often ignores nonlinear influence of inconsistency among monomers on the whole functional state under a dynamic working condition, and collaborative tracking of the charge state and the health state is difficult to realize. In addition, most models adopt a single time scale, cannot simultaneously adapt to second-level change of the state of charge and month/grade evolution of the state of health, estimation errors are easy to accumulate and amplify in complex actual working conditions, and full play of battery performance and improvement of safety management level are restricted. Disclosure of Invention Aiming at the defects of the prior art, the invention expands the estimation object from the battery monomer to the module level by establishing the functional state dynamic model and combining the double time scale estimation strategy, provides a battery module state estimation method considering the functional state dynamic characteristics, aims to effectively improve the accuracy and the dynamic response capability of the battery state estimation, and provides a data base for the long-term health management of the battery. The invention provides a battery module state estimation method considering dynamic characteristics of a functional state, which comprises the following steps: collecting characteristic parameters of the single battery, and calculating electrochemical polarization voltage and concentration polarization voltage inside the battery; respectively constructing a state-of-charge model representing short-time scale characteristics of the battery and a state-of-health model representing long-time scale characteristics of the battery according to factors influencing state estimation of the battery module; based on the electrochemical polarization voltage, the concentration polarization voltage, the state of charge model and the state of health model, a functional state dynamic model of the single battery is established; Based on the state-of-charge model and the state-of-health model, constructing a state estimation model of the single battery under a double time scale; According to the functional state dynamic model of the single battery, constructing a functional state dynamic model of the battery module; according to a state estimation model of the single battery under the double time scales and a functional state dynamic model of the battery module, constructing a state estimation model of the battery module under the double time scales; According to the state estimation model of the battery module under the double time scales, the state of charge estimation value and the state of health estimation value of the battery module at the current moment are calculated, and then according to the functional state dynamic model of the battery module, the functional state of the battery module at the current moment is calculated. Further, the characteristic parameters of the single battery comprise voltage, current, internal resistance and capacitance. Further, the specific method for respectively constructing the state-of-charge model representing the short-time scale characteristic of the battery and the state-of-health model representing the long-time scale characteristic of the battery accor