Lithium ion battery remaining service life prediction method based on AR ensemble learning model

Abstract

The invention provides a lithium ion battery remaining service life prediction method based on an AR ensemble learning model, relates to a lithium ion battery remaining service life prediction method and aims at solving the problem that an existing single AR model is limited in accuracy and low in stability during nonlinear time series prediction. The method performs prediction on the lithium ion battery remaining service life based on the AR ensemble learning model, vectors composed by input data are randomly selected to form a group of sub-vector sets by adopting a Bagging (Bootstrap Aggregating) integration method, one AR model is input into each vector set to perform parameter computation and capacity prediction, finally prediction results are integrated and output, a capacity degeneration curve and a probability density curve are drawn, and finally a final prediction output is obtained. The lithium ion battery remaining service life prediction method can improve the stability and accuracy of the lithium ion battery remaining service life prediction and is suitable for lithium ion battery remaining service life prediction.

Description

technical field

[0001] The invention relates to a method for predicting the remaining life of a lithium ion battery. Background technique

[0002] Compared with traditional nickel-cadmium or hydrogen-nickel batteries, lithium-ion batteries have the advantages of high working voltage, small size, light weight, high specific energy, long life and low self-discharge rate, etc., and become the third generation of satellites that can replace traditional batteries. energy storage power supply. If the energy storage power supply in the spacecraft uses a lithium-ion battery, it will reduce the weight of the energy storage power supply in the power subsystem from 30% to 40% to 10% to 15%, which reduces the launch cost of the spacecraft and improves the efficiency of the spacecraft. Payload.

[0003] Since the battery pack is the only energy source for the satellite during the shadow period, and the performance of the battery pack is degraded to the point that it cannot meet the nor...

Examples

specific Embodiment approach 1

[0049] Specific implementation mode one: combine figure 1 , the process of the lithium-ion battery remaining life prediction method based on the AR integrated learning model is:

[0050] Step 1: According to the capacity failure threshold of the battery, the end of life time of the battery is obtained from the capacity data of the battery, and the time quantification is represented as the number of cycles of battery charge and discharge; a% of the number of cycles at the end of life is used as the starting point for prediction; Extract capacity data, use it as the original input data F for order judgment, and standardize F to obtain standardized data Y;

[0051] Zero-meanization: calculate the mean value Fmean of the input data F, and obtain the zero-meanization sequence f=F-Fmean;

[0052] Variance standardization: Find the standard deviation σ of the sequence f f , get standardized data Y=f / σ f ;

[0053] Step 2: Calculate the 0-step autocovariance of the standardized da...

specific Embodiment approach 2

[0086] Specific implementation mode two: the specific operation steps of step 4 of this embodiment mode are:

[0087] Construct the matrix of the Yule-Wallker equation in the following form:

[0088] ρ 1 ρ 2 · · · ρ p = 1 ρ 1 · · · ρ p - 1 ...

specific Embodiment approach 3

[0092] Specific implementation mode three: the specific operation steps of step 5 of the present implementation mode are:

[0093] Step 5.1 is calculated by autocorrelation coefficient:

[0094] S=[R 0 ,R(1),R(2),R(3)] (9)

[0095] S is a vector composed of 0-3 step autocorrelation coefficients, R 0 , R(1), R(2), R(3) are 0-3 step autocorrelation coefficients respectively;

[0096] Step 5.2 calculates the Toeplitz matrix Toeplitz matrix according to S:

[0097] G=toeplitz(S) (10)

[0098] Among them, G is the Toeplitz matrix of vector S;

[0099] Step 5.3 calculates the intermediate vector W:

[0100] W=G -1 ·[R(1),R(2),R(3),R(4)] T (11)

[0101] Compute the model residual variance:

[0102] σ p 2 = 1 L 1 - p Σ t = p + ...

technical field

[0001] The invention relates to a method for predicting the remaining life of a lithium ion battery. Background technique

[0002] Compared with traditional nickel-cadmium or hydrogen-nickel batteries, lithium-ion batteries have the advantages of high working voltage, small size, light weight, high specific energy, long life and low self-discharge rate, etc., and become the third generation of satellites that can replace traditional batteries. energy storage power supply. If the energy storage power supply in the spacecraft uses a lithium-ion battery, it will reduce the weight of the energy storage power supply in the power subsystem from 30% to 40% to 10% to 15%, which reduces the launch cost of the spacecraft and improves the efficiency of the spacecraft. Payload.

[0003] Since the battery pack is the only energy source for the satellite during the shadow period, and the performance of the battery pack is degraded to the point that it cannot meet the nor...

Examples

specific Embodiment approach 1

[0049] Specific implementation mode one: combine figure 1 , the process of the lithium-ion battery remaining life prediction method based on the AR integrated learning model is:

[0050] Step 1: According to the capacity failure threshold of the battery, the end of life time of the battery is obtained from the capacity data of the battery, and the time quantification is represented as the number of cycles of battery charge and discharge; a% of the number of cycles at the end of life is used as the starting point for prediction; Extract capacity data, use it as the original input data F for order judgment, and standardize F to obtain standardized data Y;

[0051] Zero-meanization: calculate the mean value Fmean of the input data F, and obtain the zero-meanization sequence f=F-Fmean;

[0052] Variance standardization: Find the standard deviation σ of the sequence f f , get standardized data Y=f / σ f ;

[0053] Step 2: Calculate the 0-step autocovariance of the standardized da...

specific Embodiment approach 2

[0086] Specific implementation mode two: the specific operation steps of step 4 of this embodiment mode are:

[0087] Construct the matrix of the Yule-Wallker equation in the following form:

[0088] ρ 1 ρ 2 · · · ρ p = 1 ρ 1 · · · ρ p - 1 ...

specific Embodiment approach 3

[0092] Specific implementation mode three: the specific operation steps of step 5 of the present implementation mode are:

[0093] Step 5.1 is calculated by autocorrelation coefficient:

[0094] S=[R 0 ,R(1),R(2),R(3)] (9)

[0095] S is a vector composed of 0-3 step autocorrelation coefficients, R 0 , R(1), R(2), R(3) are 0-3 step autocorrelation coefficients respectively;

[0096] Step 5.2 calculates the Toeplitz matrix Toeplitz matrix according to S:

[0097] G=toeplitz(S) (10)

[0098] Among them, G is the Toeplitz matrix of vector S;

[0099] Step 5.3 calculates the intermediate vector W:

[0100] W=G -1 ·[R(1),R(2),R(3),R(4)] T (11)

[0101] Compute the model residual variance:

[0102] σ p 2 = 1 L 1 - p Σ t = p + ...

Images