Method for estimating health state of lithium ion battery under different fast charging working conditions

Abstract

The invention discloses a method for estimating the health state of a lithium ion battery under different fast charging working conditions, and the method comprises the steps: obtaining a whole-process voltage curve of the lithium ion battery under different fast charging working conditions, and carrying out the slicing of the whole-process voltage curve based on sampling points; performing sample entropy calculation on the sliced voltage curve segments to obtain sample entropy data of the voltage curve segments; obtaining the capacity data of the lithium ion battery, carrying out correlation coefficient calculation on the sample entropy data and the capacity data, and selecting the sample entropy data with the highest correlation based on the value of the correlation coefficient; and constructing a circulating network model, taking the sample entropy data with the highest correlation as input data of the circulating network model, and predicting to obtain the health state of the lithium ion battery. According to the method, the self-adaptive battery health state estimation model is obtained by utilizing the historical information of the lithium ion battery under different fast charging working conditions, and the modeling training requirement on a new data set is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of lithium-ion battery state estimation, and in particular relates to a method for estimating the health state of a lithium-ion battery under different fast charging conditions. Background Art

[0002] Lithium-ion batteries have the characteristics of high energy density, long life, low self-discharge rate, and rapid cost reduction in recent years. Lithium-ion batteries have become the main energy storage components in many industrial fields, from portable electronics to electric vehicles and power grids. By accurately estimating the state of health (SOH) of the battery, the health of the battery can be evaluated in time, potential failures can be prevented, and optimal performance and cost-effectiveness can be achieved under the premise of ensuring safety, reliability and durability. The state of health of the battery is usually defined as the ratio of the current maximum capacity to the nominal capacity of a new battery. However, using current commercial sensing technology, the battery capacity cannot be directly physically measured.

[0003] Multiple degradation mechanisms may occur throughout the battery's lifetime, such as electrode fatigue and stratification, solid electrolyte interface growth, electrolyte decomposition, etc. In this context, it is difficult to build a battery degradation model that involves all these nonlinear and coupled dynamics. In current data-driven approaches, there are high requirements on the availability and quality of data. Because battery management systems can only measure battery terminal voltage, current, and surface temperature, the availability of datasets covering various operating conditions, aging stress factors, and failure modes is critical for accurate estimation of battery health status. Aging stresses and aging trajectories are uncertain in the load and temperature distribution among different cells. For example, the charging current amplitude is not uniform in various fast charging strategies.

[0004] Therefore, there is currently a lack of an efficient method that can better utilize historical information to obtain an adaptive battery health status estimation model and reduce the need for modeling training of new data sets. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a health status estimation method for lithium-ion batteries under different fast charging conditions to solve the problems existing in the above-mentioned prior art.

[0006] To achieve the above object, the present invention provides a method for estimating the health state of a lithium-ion battery under different fast charging conditions, comprising:

[0007] Obtaining a full-process voltage curve of a lithium-ion battery under different fast-charging conditions, and slicing the full-process voltage curve based on sampling points;

[0008] Performing sample entropy calculation on the sliced ​​voltage curve segment to obtain sample entropy data of the voltage curve segment;

[0009] Acquire capacity data of the lithium-ion battery, calculate a correlation coefficient between the sample entropy data and the capacity data, and select the sample entropy data with the highest correlation based on the value of the correlation coefficient;

[0010] A cyclic network model is constructed, and the sample entropy data with the highest correlation is used as input data of the cyclic network model to predict the health status of the lithium-ion battery.

[0011] Preferably, the sliced ​​voltage curve segments are N-dimensional time series sampled at equal time intervals.

[0012] Preferably, the process of calculating the sample entropy of the sliced ​​voltage curve segment is as follows:

[0013] Based on the N-dimensional time series, the m-dimensional vector X is reconstructed. m (1),X m (2),…,X m (n-m+1), where X m (i)=[x(i),x(i+1),…,x(i+m-1)];

[0014] Calculate two corresponding elements X m (i) With X m (j) the absolute value of the maximum difference between m (i),X m (j)];

[0015] Set the threshold r, and count d[X m (i),X m (j)] is less than the ratio of the number of r to the total number of distances Nm m i (r);

[0016] Find the ratio B i m (r) The average value B for all values ​​of i m (r);

[0017] Let k = m + 1, repeatedly calculate the ratio and the average value, and obtain the first parameter A k (r);

[0018] Based on the first parameter A k (r) and the average value B m (r), the sample entropy data of the voltage curve segment is calculated.

[0019] Preferably, the formula for calculating the correlation coefficient between the sample entropy data and the capacity data is:

[0020]

[0021] Where ρ is the Pearson correlation coefficient, X is the sample entropy sequence, X = {x i} N i＝1 , Y is the capacity sequence, Y={y i} N i-1 , and are the means of X and Y respectively.

[0022] Preferably, the recurrent network model comprises: a bidirectional long short-term memory network and an attention layer;

[0023] The bidirectional long short-term memory network is two long short-term memory networks, one of which processes the input data in a forward direction to obtain a first output, and the other processes the input data in a reverse direction to obtain a second output, and the first output and the second output are connected to obtain an output vector of the bidirectional long short-term memory network;

[0024] Based on the weight of the attention layer, the output vectors are weighted summed to generate a context vector.

[0025] Preferably, the long short-term memory network comprises: a forget gate, an input gate and an output gate;

[0026] The calculation formula of the forget gate is:

[0027] f t =σ(W fh h t-1 +W fx x t +b f )

[0028] Among them, σ is the activation function of the forget gate, W fh is the weight of fh, b f is the bias of f;

[0029] The calculation formula of the input gate is:

[0030] i t =σ(W ih h t-1 +W ix x t +b i )

[0031]

[0032] Among them, W ih , W ix , W Ch and W Cxrepresents the weight of the corresponding layer, b i and b f is the bias, is the candidate state of the current unit;

[0033]

[0034] Among them, C t-1 is the previous internal state, C t is the current state;

[0035] Use sigmoid layer o t and tanh layer h t To generate the output value of the long short-term memory network, the formula is:

[0036] o t =σ(W oh h t-1 +W ox x t +b o )

[0037] h t =o t tanh(C t ).

[0038] Preferably, the formula for generating the context vector is:

[0039]

[0040] Among them, c is the context vector, α t,i is the combination assigned to input position i and output position t, v a and W a is the weight matrix that needs to be learned.

[0041] Preferably, the method further comprises: training the recurrent network model based on fine-tuning transfer learning, first freezing the weight coefficients of the initial seven layers of the recurrent network model, and secondly modifying the weight coefficients of the last four layers of the recurrent network model.

[0042] Compared with the prior art, the present invention has the following advantages and technical effects:

[0043] The present invention provides a method for estimating the health state of a lithium-ion battery under different fast-charging conditions, comprising: firstly obtaining a full-process voltage curve of a lithium-ion battery, and slicing the full-process voltage curve based on sampling points; then performing sample entropy calculation on the sliced ​​voltage curve segments to obtain sample entropy data of the voltage curve segments; then obtaining capacity data of the lithium-ion battery, performing correlation coefficient calculation on the sample entropy data and the capacity data, and selecting sample entropy data with the highest correlation based on the value of the correlation coefficient; finally, constructing a cyclic network model, using the sample entropy data with the highest correlation as input data of the cyclic network model, and predicting the health state of the lithium-ion battery.

[0044] The present invention firstly takes into account various non-standardized fast charging protocols in the current market, and uses feature engineering to extract feature information that is related to battery aging and has strong robustness under different fast charging strategies; then, targeting the temporal characteristics of battery aging information, a method using a deep recurrent network structure to capture the long-term dependency information of battery aging time series data is proposed. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0046] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0047] Figure 2 This is a voltage and current curve diagram of the entire charging and discharging process of a battery with a one-step fast charging strategy according to an example of an embodiment of the present invention;

[0048] Figure 3 The voltage and current curve diagram of the whole charging and discharging process of a battery with a two-step fast charging strategy according to an example of an embodiment of the present invention;

[0049] Figure 4 A schematic diagram of a voltage curve of the entire charging and discharging process of a battery in its entire life cycle according to an embodiment of the present invention;

[0050] Figure 5 A schematic diagram of the first voltage curve segment of the full-life charge and discharge process of a battery according to an embodiment of the present invention;

[0051] Figure 6 A schematic diagram of a second voltage curve segment during the full life charge and discharge process of a battery according to an embodiment of the present invention;

[0052] Figure 7 It is a schematic diagram of the third voltage curve segment of the full-life charge and discharge process of the battery according to an embodiment of the present invention, that is, a voltage curve segment of the CC charge and CC-CV charge conversion stage;

[0053] Figure 8 It is a schematic diagram of the fourth voltage curve segment of the full-life charge and discharge process of the battery according to an embodiment of the present invention;

[0054] Fig. 9 It is a schematic diagram of the fifth voltage curve segment of the full-life charge and discharge process of the battery according to an embodiment of the present invention;

[0055] Fig.10 It is a schematic diagram of the sixth voltage curve segment of the full-life charge and discharge process of the battery according to an embodiment of the present invention;

[0056] Fig.11 It is a schematic diagram of the seventh voltage curve segment of the full-life charge and discharge process of the battery according to an embodiment of the present invention;

[0057] Fig.12 A heat map of correlation coefficients between sample entropy of seven voltage slice segments and battery capacity according to an embodiment of the present invention;

[0058] Fig.13 A block diagram of a BiLSTM-Att neural network according to an embodiment of the present invention;

[0059] Fig.14 This is a comparison chart of SOH prediction results of test battery 1 after transfer learning and without transfer learning according to an embodiment of the present invention;

[0060] Fig.15 This is a comparison chart of the SOH prediction results of the test battery 2 after transfer learning and without transfer learning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0061] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0062] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0063] Embodiment 1

[0064] This embodiment provides a method for estimating the health status of a lithium-ion battery under different fast charging conditions, such as Figure 1 As shown, including:

[0065] Step 1: Based on the full life cycle charge and discharge voltage data of lithium-ion batteries under different fast charging conditions, slice the voltage curve of the whole process based on the sampling points.

[0066] The voltage and current curves of the battery during the entire charging and discharging process under the one-step fast charging strategy and the two-step fast charging strategy are as follows: Figure 2-Figure 3 As shown, the difference between the two operating conditions lies in the number of CC charging stages.

[0067] Figure 4 It is the voltage curve of the whole process of battery charge and discharge throughout its life. Each curve represents a whole process of charge and discharge. The number of curves is equal to the number of charge and discharge cycles of the battery throughout its life. This battery is used as an example for slicing. Specifically, it is known that the voltage curve of any cycle is V(t), and the sampling point time of the last constant current (CC) charging and constant current-constant voltage (CC-CV) charging conversion stage is t0. Then V(t), t∈(t0-T, t0+T) is the voltage curve segment of the conversion stage, and its length is 2T. In addition, the remaining voltage curves are sliced ​​from beginning to end according to the length of 2T, and the curves with less than 2T at the end are discarded to obtain num voltage curve segments. These num segments are used as the control group, and a total of num+1 voltage segments are obtained. Each curve is sliced ​​separately, and the results are as follows Figure 5-Figure 11 shown.

[0068] Step 2: Based on the sliced ​​voltage curve segments, use the sample entropy algorithm to calculate respectively.

[0069] The sample entropy algorithm is as follows:

[0070] S21. The voltage curve segment is an N-dimensional time series D = {x1, x2, ..., xN} obtained by sampling at equal time intervals;

[0071] S22. Reconstruct the m-dimensional vector X m (1),X m (2),…,X m (n-m+1), where X m (i)=[x(i),x(i+1),…,x(i+m-1)];

[0072] S23.Define X m (i) With X m (j) The distance d[X m (i),X m (j)] is the absolute value of the maximum difference between two corresponding elements, that is,

[0073] d[X m (i),X m (j)]=max[|x(i+k)-x(j+k)|];

[0074] S24. Given a threshold r, for each i value, statistics d[X m (i),X m The ratio of the number of (j)] less than r to the total number of distances Nm is denoted as Bm i (r), that is

[0075]

[0076] S25.Seek B i m (r) The average value of all i values, denoted as B m (r), i.e.

[0077]

[0078] S26. Let k = m + 1, repeat steps S24-S25, and obtain:

[0079]

[0080] S27. Sample entropy (sampEn) is

[0081]

[0082] Since N cannot be ∞ in actual computing applications, when N takes a finite value, the sample entropy is approximately

[0083] SampEn=—ln[A k (r) / B m (r)];

[0084] Step 3: Based on the sample entropy and capacity data of each voltage curve segment, calculate the Pearson correlation coefficient respectively.

[0085] For two given sequences, the sample entropy sequence X = {xi} N i=1,capacity sequence Y={yi} N i-1, the Pearson correlation coefficient ρ(X,Y) can be calculated as:

[0086]

[0087] Where ρ is the Pearson correlation coefficient. and are the means of X and Y respectively. The sample entropy data with the highest correlation is used as the input feature. The correlation coefficient between the voltage segment sample entropy and the battery capacity is shown in Figure 4 As shown, HI1-7 represent the sample entropy features of 7 voltage segments respectively.

[0088] Step 4: Build the BiLSTM-Att network structure.

[0089] The long short-term memory network LSTM is mainly composed of a forget gate, an input gate, and an output gate. The forget gate determines whether the data from the current unit should be discarded:

[0090] ft=σ(Wfhht -1 +Wfxxt+b f )

[0091] Where σ is the activation function of the forget gate. fh is the weight of fh, b f is the bias of f.

[0092] The input gate decides which new data should be updated. The determination of new information is activated by a sigmoid layer. Then, the tanh layer generates a vector to store the candidate state of the current unit. It can be expressed as:

[0093] i t =σ(W ih h t-1 +W ix x t +b i )

[0094]

[0095] Where W ih , W ix , W Ch and W Cx represents the weight of the corresponding layer, b i and b f is bias.

[0096] The above gate can be used to convert the previous internal state C t-1 Update to current status C t .

[0097]

[0098] Finally, a sigmoid layer is used t and tanh layer h t To generate the output value of LSTM, it can be expressed as:

[0099] o t =σ(W oh h t-1 +W ox x t +b o )

[0100] h t =o t tanh(C t )

[0101] The Bidirectional LSTM (BiLSTM) network consists of two LSTM components. One processes the input sequence in the forward direction, while the other processes it in the reverse direction. After processing, the outputs of the two LSTMs are concatenated to obtain the final BiLSTM output. Consider the input sequence X = {x i} n i＝1 To the target sequence Y = {x i} n i＝1 The mapping task is n, where n is the length of the sequence. The encoder state can be defined based on the hidden layer of BiLSTM:

[0102]

[0103] in, and They represent the forward hidden state and reverse hidden state of BiLSTM respectively.

[0104] The weights calculated by the Attention layer are used to perform weighted summation on the output vectors of the BiLSTM layer. The specific network structure is as follows: Fig.13 This generates a context vector, which is a weighted representation of all elements in the input sequence. The decoder network contains the hidden state s of the output at position t. t =f(s t-1 ,y t -1,c t ), where the context vector c is the sum of the hidden states of the input sequence.

[0105]

[0106] α t,i The score assigned to the combination of input position i and output position t depends on the degree of match between them. t ,i} is a set of weights indicating how many source hidden states should be considered for each output.

[0107] α is parameterized by a single hidden layer feed-forward neural network that is trained together with the rest of the model. Therefore, the score function takes the following format (tanh is used as the nonlinear activation function):

[0108]

[0109] Among them, v a and W a is the weight matrix that needs to be learned.

[0110] Step 5: Fine-tune transfer learning based on the pre-trained model.

[0111] The BiLSTM-Att model structure constructed in this embodiment is shown in Table 1, with a total of 11 layers. The pre-trained network used is relatively complex, so it takes a lot of time to train from scratch. The fine-tuning strategy adopted in this embodiment includes freezing the weight coefficients of the initial seven layers of the network. During the migration training process, only the weight coefficients of the last four layers of the network are modified.

[0112] Table 1

[0113]

[0114] Based on the one-step fast charging strategy data and the two-step fast charging strategy data, the model is trained and tested respectively to verify the accuracy of the improved deep cycle network proposed in this embodiment in predicting the battery health status. Two battery data are taken from each of the two fast charging conditions and four network models are used for verification. The SOH prediction results are quantified by four evaluation indicators, as shown in Table 2. Among them, MAE is the mean absolute error loss, MAPE is the mean absolute percentage error, RMSE is the root mean square error;

[0115] Table 2

[0116]

[0117] Based on the pre-trained model of the one-step fast charging strategy data, fine-tune the transfer learning, verify the number of fine-tuning layers through experiments, and freeze the remaining network layers;

[0118] Based on the fine-tuned migration model and two-step fast charging strategy data, fine-tuned migration learning training was performed and tested on the two-step fast charging strategy data test set to verify the accuracy, efficiency and robustness of the migration learning method proposed in this embodiment in predicting the health status of batteries across charging conditions. Two test batteries under two-step fast charging conditions were randomly selected for migration model and non-migration model verification. The results are shown in Figure 2. Figure 14-15 The training time is shown in Table 3, where BiLSTM-Att-TL is a migration model and the other is a non-migrated model.

[0119] Table 3

[0120]

[0121] This embodiment proposes a transfer learning algorithm to process knowledge migration between different fast charging modes to reduce the data requirements and model training time cost of different fast charging modes.

[0122] Beneficial effects of this embodiment:

[0123] This embodiment proposes a method for estimating the health status of lithium-ion batteries under different fast-charging conditions. It can extract high robustness features related to battery aging based on partial voltage curves of batteries under different fast-charging strategies; it can accurately predict the battery health status using battery aging features; it can accurately predict the battery health status by using transfer learning of pre-trained models for batteries under different fast-charging conditions. The transfer learning method proposed in this embodiment solves the problem of small data volume of different fast-charging data sets, and reduces the time required for model training by 64%, saving computing resources.

[0124] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for estimating the health status of a lithium-ion battery under different fast charging conditions, characterized in that: The following steps are involved: Obtaining a full-process voltage curve of a lithium-ion battery under different fast-charging conditions, and slicing the full-process voltage curve based on sampling points; Performing sample entropy calculation on the sliced ​​voltage curve segment to obtain sample entropy data of the voltage curve segment; Acquire capacity data of the lithium-ion battery, calculate a correlation coefficient between the sample entropy data and the capacity data, and select the sample entropy data with the highest correlation based on the value of the correlation coefficient; A cyclic network model is constructed, and the sample entropy data with the highest correlation is used as input data of the cyclic network model to predict the health status of the lithium-ion battery.

2. The method for estimating the health status of a lithium-ion battery under different fast-charging conditions according to claim 1, characterized in that: The sliced ​​voltage curve segments are N-dimensional time series sampled at equal time intervals.

3. The method for estimating the health status of a lithium-ion battery under different fast charging conditions according to claim 2, characterized in that: The process of calculating sample entropy for the sliced ​​voltage curve segment: Based on the N-dimensional time series, the m-dimensional vector X is reconstructed. m (1),X m (2),…,X m (n-m+1), where X m (i)=[x(i),x(i+1),…,x(i+m-1)]; Calculate two corresponding elements X m (i) With X m (j) the absolute value of the maximum difference between m (i),X m (j)]; Set the threshold r, and count d[X m (i),X m (j)] is less than the ratio of the number of r to the total number of distances Nm m i (r); Find the ratio B i m (r) The average value B for all values ​​of i m (r); Let k = m + 1, repeatedly calculate the ratio and the average value, and obtain the first parameter A k (r); Based on the first parameter A k (r) and the average value B m (r), the sample entropy data of the voltage curve segment is calculated.

4. The method for estimating the health status of a lithium-ion battery under different fast charging conditions according to claim 1, characterized in that: The formula for calculating the correlation coefficient between the sample entropy data and the capacity data is: Where ρ is the Pearson correlation coefficient, X is the sample entropy sequence, X = {x i } N i＝1 , Y is the capacity sequence, Y={y i } N i-1 , and are the means of X and Y respectively.

5. The method for estimating the health status of a lithium-ion battery under different fast-charging conditions according to claim 1, characterized in that: The recurrent network model includes: a bidirectional long short-term memory network and an attention layer; The bidirectional long short-term memory network is two long short-term memory networks, one of which processes the input data in a forward direction to obtain a first output, and the other processes the input data in a reverse direction to obtain a second output, and the first output and the second output are connected to obtain an output vector of the bidirectional long short-term memory network; Based on the weight of the attention layer, the output vectors are weighted summed to generate a context vector.

6. The method for estimating the health status of a lithium-ion battery under different fast-charging conditions according to claim 5, characterized in that: The long short-term memory network includes: a forget gate, an input gate and an output gate; The calculation formula of the forget gate is: f t =σ(W fh h t-1 +W fx x t +b f ) Among them, σ is the activation function of the forget gate, W fh is the weight of fh, b f is the bias of f; The calculation formula of the input gate is: i t =σ(W ih h t-1 +W ix x t +b i ) Among them, W ih , W ix , W Ch and W Cx represents the weight of the corresponding layer, b i and b f is the bias, is the candidate state of the current unit; Among them, C t-1 is the previous internal state, C t is the current state; Use sigmoid layer o t and tanh layer h t To generate the output value of the long short-term memory network, the formula is: o t =σ(W oh h t-1 +W ox x t +b o ) h t =o t fishy(C) t )。 7. The method for estimating the health status of a lithium-ion battery under different fast-charging conditions according to claim 5, characterized in that: The formula for generating the context vector is: Among them, c is the context vector, α t,i is the combination assigned to input position i and output position t, v a and W a is the weight matrix that needs to be learned.

8. The method for estimating the health status of a lithium-ion battery under different fast-charging conditions according to claim 1, characterized in that: Also includes: The recurrent network model is trained based on fine-tuning transfer learning. First, the weight coefficients of the initial seven layers of the recurrent network model are frozen, and then the weight coefficients of the last four layers of the recurrent network model are modified.

Images