Method and system for identifying electrochemical model parameters based on capacity change rate

1 . A method for improving accuracy of lithium battery state estimation by identifying electrochemical model parameters based on a capacity change rate, comprising:

acquiring physical operating data from an actual operation process of a lithium battery, wherein the lithium battery has remained idle for a period longer than a preset time length to ensure an internal equilibrium state; and cleaning an actual operating data set from the physical operating data to include an SOC change interval greater than a preset threshold value;

generating a simulated operating data set through simulation of a preset electrochemical model representative of the lithium battery;

performing parameter identification of the electrochemical model based on a preset first loss function calculated between the actual operating data set and the simulated operating data set;

calculating a capacity convergence coefficient between a battery capacity obtained through simulation according to a historical parameter identification result and a battery capacity obtained through simulation according to a current parameter identification result;

comparing the capacity convergence coefficient with a preset convergence threshold value indicative of normal battery aging conditions;

when the capacity convergence coefficient is greater than the preset convergence threshold value, discarding the current parameter identification result and regenerating the simulated operating data set using updated electrochemical model parameters; and

when the capacity convergence coefficient is not greater than the preset convergence threshold value, outputting an electrochemical model parameter set as a parameter identification result for use in monitoring the lithium battery.

2 . The method of claim 1 , wherein after the generating the simulated operating data set through simulation of the preset electrochemical model and before the comparing the capacity convergence coefficient with the preset convergence threshold value, the method further comprises:

based on the preset first loss function, calculating a first loss function result value between the actual operating data set and the simulated operating data set;

when the first loss function result value is greater than a preset first loss function result threshold value, storing the current parameter identification result in a historical parameter identification result database, and regenerating the simulated operating data set and the parameter identification result; and

when the first loss function result value is not greater than the preset first loss function result threshold value, calculating a capacity convergence coefficient between the battery capacity obtained through simulation according to the previous historical parameter identification result in the historical parameter identification result database and the battery capacity obtained through simulation according to the current parameter identification result.

3 . The method of claim 2 , wherein the first loss function comprises a voltage mean square error loss function, and the calculating the first loss function result value between the actual operating data set and the simulated operating data set based on the preset first loss function comprises:

calculating a voltage mean square error between the actual operating data set and the simulated operating data set as the first loss function result value, wherein a formula is as follows:

MSE

=

1

N

⁢

∑

i

=

1

N

⁢

(

V

s

⁢

i

⁢

m

,

i

-

V

real

,

i

)

2

;

wherein V sim,i is a model simulation output voltage value of the ith sampling point in the simulated operating data set, V real,i is a measured voltage value of the ith sampling point of the actual operating data set, and N is a number of voltage data points.

4 . The method of claim 1 , wherein the calculating the capacity convergence coefficient between the battery capacity obtained through simulation according to the historical parameter identification result and the battery capacity obtained through simulation according to the current parameter identification result comprises:

after conducting parameter standardization on the historical parameter identification result, calculating, using the electrochemical model, the battery capacity obtained through simulation according to the historical parameter identification result;

after conducting parameter standardization on the current parameter identification result, calculating, using the electrochemical model, the battery capacity obtained through simulation according to the current parameter identification result; and

calculating the capacity convergence coefficient between the battery capacity obtained through simulation according to the historical parameter identification result and the battery capacity obtained through simulation according to the current parameter identification result, wherein a formula is as follows:

δ

=

❘

"\[LeftBracketingBar]"

C

max

,

k

-

C

max

,

k

-

1

C

max

,

k

-

1

❘

"\[RightBracketingBar]"

;

wherein δ is the capacity convergence coefficient, C max,k-1 is the battery capacity obtained through simulation according to the historical parameter identification result, and C max,k is the battery capacity obtained through simulation according to the current parameter identification result.

5 . The method of claim 1 , wherein the method further comprises optimizing generalization capability of the electrochemical model parameter set by:

when the capacity convergence coefficient is not greater than the convergence threshold value, physically aligning the model state by adjusting a solid-phase initial concentration in the simulated operating data set to be the same as a solid-phase initial concentration derived from the actual physical state of the lithium battery in the actual operating data set;

performing a secondary validation based on a preset second loss function by calculating a second loss function result value between the actual operating data set and the simulated operating data set after adjusting the solid-phase initial concentration;

when the second loss function result value is greater than a preset second loss function result threshold value, identifying a mismatch between the electrochemical model and actual battery state and regenerating the simulated operating data set; and

when the second loss function result value is not greater than the preset second loss function result threshold value, outputting the electrochemical model parameter set as the parameter identification result to calibrate a battery management system for accurate power station operation.

6 . The method of claim 5 , wherein the second loss function comprises a voltage difference loss function, and the based on the preset second loss function, calculating the second loss function result value between the actual operating data set and the simulated operating data set after adjusting the solid-phase initial concentration comprises:

calculating a voltage difference value between a first voltage value output from the actual operating data set and a first voltage value output from the simulated operating data set after adjusting the solid-phase initial concentration, and taking the voltage difference value as the second loss function result value, wherein a formula is as follows:

| V sim,1 −V real,1 |;

wherein V sim,1 is a model simulation output first voltage value from the simulated operating data set, and V real,1 is a measured first voltage value from the actual operating data set for the lithium battery.

7 . The method of claim 5 , wherein the second loss function further comprises a voltage mean square error loss function.

8 . A non-transitory tangible computer-readable storage medium on which at least one instruction is stored, wherein the at least one instruction is loaded and executed by a processor to implement operations of the method for improving accuracy of lithium battery state estimation according to claim 1 .

9 . A system for improving accuracy of lithium battery state estimation by identifying electrochemical model parameters based on a capacity change rate, comprising:

an acquisition module, configured to acquire physical operating data from an actual operation process of a lithium battery, wherein the lithium battery has remained idle for a period longer than a preset time length to ensure an internal equilibrium state; and clean an actual operating data set from the physical operating data to include an SOC change interval greater than a preset threshold value;

a simulation module, configured to generate a simulated operating data set through simulation of a preset electrochemical model representative of the lithium battery;

an identification module, respectively connected with the acquisition module and the simulation module, configured to perform parameter identification of the electrochemical model based on a preset first loss function calculated between the actual operating data set and the simulated operating data set;

a calculation module, respectively connected with the identification module and the simulation module, configured to calculate a capacity convergence coefficient between a battery capacity obtained through simulation according to a historical parameter identification result and a battery capacity obtained through simulation according to a current parameter identification result;

a comparison module, connected with the calculation module, configured to compare the capacity convergence coefficient with a preset convergence threshold value indicative of normal battery aging conditions;

a generation module, connected with the comparison module, configured to discard the current parameter identification result and regenerate the simulated operating data set using updated electrochemical model parameters when the capacity convergence coefficient is greater than the preset convergence threshold value; and

an output module, connected with the comparison module, configured to output an electrochemical model parameter set as a parameter identification result for use in monitoring the lithium battery when the capacity convergence coefficient is not greater than the preset convergence threshold value.