Battery state of health prediction method and apparatus

Abstract

The application provides a battery state of health prediction method and device, the method comprises the following steps: performing a plurality of battery cycle life experiments on a target type of test battery to obtain at least one health factor sequence of the test battery under the same battery state of health sequence; according to each health factor sequence, the objective weight method is used to determine the weight value corresponding to each health factor; according to the weight value, the initial weight value of the neural network prediction model is determined, and each health factor sequence is input into the neural network prediction model as a training sample for prediction to obtain the corresponding prediction value; based on the deviation between the prediction value and each sequence value in the battery state of health sequence, the neural network prediction model is trained to obtain a trained neural network prediction model; health factor data of the target battery at the current time is obtained; the health factor data is input into the trained neural network prediction model to predict the battery state of health of the target battery, and accurate monitoring of the battery state of health is realized.

Description

Technical Field

[0001] This invention relates to the field of battery monitoring technology, and in particular to a method and apparatus for predicting battery health status. Background Technology

[0002] The working environment in coal mines is unique, with high humidity and dust levels, presenting numerous environmental factors that are detrimental to the safety performance of lithium batteries. Furthermore, thermal runaway of lithium batteries can lead to greater economic losses and even injuries or fatalities. Therefore, monitoring the health status of lithium batteries is essential. Summary of the Invention

[0003] The present invention aims to at least partially solve one of the technical problems in the related art.

[0004] Therefore, the first objective of this invention is to propose a battery health state prediction method, which obtains at least one health factor sequence of the target battery under the same battery health state sequence by conducting multiple battery cycle life experiments on the target battery, and trains a neural network prediction model based on the obtained health factor sequence, thereby improving the training efficiency and prediction accuracy of the neural network prediction model and realizing accurate monitoring of battery health state.

[0005] The second objective of this invention is to provide a battery health status prediction device.

[0006] The third objective of this invention is to provide an electronic device.

[0007] The fourth objective of this invention is to provide a computer-readable storage medium.

[0008] The fifth objective of this invention is to provide a computer program product.

[0009] To achieve the above objectives, a first aspect of the present invention provides a method for predicting battery health status, comprising:

[0010] Multiple battery cycle life experiments were conducted on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence.

[0011] Based on the health factor sequences, the weight value corresponding to each health factor is determined using the objective weighting method.

[0012] Based on the weight values, the initial weight values ​​of the neural network prediction model are determined, and each of the health factor sequences is used as a training sample and input into the neural network prediction model to make a prediction, so as to obtain the corresponding predicted value.

[0013] Based on the deviation between the predicted value and each sequence value in the battery health state sequence, the neural network prediction model is trained to obtain a trained neural network prediction model.

[0014] Obtain the current health factor data of the target battery, wherein the target battery is a battery of the target type;

[0015] The health factor data is input into the trained neural network prediction model to predict the battery health status of the target battery.

[0016] To achieve the above objectives, a second aspect of the present invention provides a battery health status prediction device, comprising:

[0017] The first processing module is used to perform multiple battery cycle life experiments on a target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence.

[0018] The determination module is used to determine the weight value corresponding to each health factor based on the health factor sequence using an objective weighting method;

[0019] The second processing module is used to determine the initial weight values ​​of the neural network prediction model based on the weight values, and to input each of the health factor sequences as training samples into the neural network prediction model for prediction in order to obtain the corresponding predicted values.

[0020] The training module is used to train the neural network prediction model based on the deviation between the predicted value and each sequence value in the battery health state sequence, so as to obtain a trained neural network prediction model.

[0021] The acquisition module is used to acquire the health factor data of the target battery at the current moment, wherein the target battery is a battery of the target type;

[0022] The prediction module is used to input the health factor data into the trained neural network prediction model to predict the battery health status of the target battery.

[0023] To achieve the above objectives, a third aspect of the present invention provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the battery health state prediction method of the first aspect described above.

[0024] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium storing computer instructions for causing the computer to execute the battery health state prediction method of the first aspect.

[0025] To achieve the above objectives, a fifth aspect of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the battery health state prediction method of the first aspect.

[0026] The technical solutions provided by the embodiments of the present invention have the following beneficial effects:

[0027] By conducting multiple battery cycle life experiments on target-type test batteries, at least one health factor sequence is obtained for each test battery under the same battery health state sequence. Based on each health factor sequence, an objective weighting method is used to determine the weight value corresponding to each health factor. This weight value is then used to determine the initial weight values ​​of the neural network prediction model. Each health factor sequence is used as training samples and input into the neural network prediction model to obtain the corresponding predicted values. The neural network prediction model is then trained based on the deviation between the predicted values ​​and the values ​​of each sequence in the battery health state sequence. After acquiring the current health factor data of the target battery, this data is input into the trained neural network prediction model to predict the battery health state. This enables accurate monitoring of battery health state. Furthermore, since the initial weight values ​​of the neural network prediction model can be determined based on the weight values ​​corresponding to each health factor using the objective weighting method, the training efficiency of the neural network prediction model can be effectively improved.

[0028] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0029] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0030] Figure 1 This is a flowchart illustrating a battery health status prediction method provided in an embodiment of the present invention.

[0031] Figure 2 This is a schematic diagram of a health factor in a given scenario, provided by an embodiment of the present invention.

[0032] Figure 3 This is a schematic diagram of a BP neural network model in a specific scenario provided by an embodiment of the present invention;

[0033] Figure 4 This is a flowchart illustrating another battery health status prediction method provided in an embodiment of the present invention.

[0034] Figure 5 This is a flowchart illustrating another battery health status prediction method provided in an embodiment of the present invention.

[0035] Figure 6 This is a schematic diagram of the structure of a battery health status prediction device provided in an embodiment of the present invention;

[0036] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0037] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0038] The battery health status prediction method and apparatus of the present invention are described below with reference to the accompanying drawings.

[0039] Figure 1 This is a flowchart illustrating a battery health status prediction method provided in an embodiment of the present invention.

[0040] It should be noted that the battery health status prediction method in this embodiment is executed by a battery health status prediction device, which can be implemented by software and / or hardware, and can be configured in an electronic device. The following description takes an electronic device as the execution subject.

[0041] like Figure 1 As shown, the battery health status prediction method includes the following steps:

[0042] Step 101: Perform multiple battery cycle life experiments on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence.

[0043] Among them, health factors refer to variables that characterize the health status of a battery. For example, the health factors of a lithium battery refer to variables that characterize the health status of a lithium-ion battery. Based on the actual working environment in a mine, factors such as temperature, humidity, and pressure can be analyzed.

[0044] Among them, at least one health factor sequence includes one or more of temperature sequence, humidity sequence, and battery expansion force sequence. In this case, the corresponding health factor is as follows: Figure 2 As shown, this includes one or more of the following: ambient temperature, ambient humidity, and battery expansion force.

[0045] The standard definition of battery health status is the ratio of the capacity released by a power battery from a fully charged state to the cutoff voltage under standard conditions to its corresponding nominal capacity (actual initial capacity). This ratio reflects the battery's health status. Simply put, it's the ratio of the actual value to the nominal value of certain directly measurable or indirectly calculated performance parameters after a period of battery use. It's used to determine the battery's declining health status and measure its degree of health.

[0046] Accurate health assessment can provide a full understanding of the battery's current condition, enabling the development of maintenance plans based on underlying assumptions, correction of various performance parameters, reduction or avoidance of risks, or maintenance and replacement of individual cells whose performance does not meet usage requirements, thereby reducing operating costs.

[0047] Battery cycle life refers to the number of cycles a battery can withstand before its capacity drops to a specified value under a certain charge-discharge regime. The number of cycles refers to the number of complete charge-discharge cycles a battery undergoes, which can be estimated by the actual discharge capacity and the design capacity. Each time the accumulated discharge capacity equals the design capacity, one cycle is counted.

[0048] As an example, suppose a battery has a cycle life of 500 cycles, meaning it can be used for 500 charge-discharge cycles. One charge-discharge cycle is not the same as one charge. Let's assume one charge-discharge cycle refers to the process of using all the battery's charge from full to empty, and then charging it back to full. If a battery starts at full charge, uses one-tenth of its capacity, and then is fully charged again, that's one-tenth of a cycle. Ten such cycles would essentially constitute one complete cycle.

[0049] In some embodiments, at least one health factor sequence of the test battery under the same battery health state sequence can be obtained by performing multiple battery cycle life experiments on the test battery of the target type. For example, assuming the battery health state sequence is X = {x(k)|k = 1, 2, ..., n}, then it is necessary to measure at least one health factor sequence corresponding to the battery health states x(1), x(2), ..., x(n). Assuming that the measured at least one health factor sequence includes a temperature sequence, a humidity sequence, and a battery expansion force sequence, wherein the temperature sequence is Y1 = {y1(k)|k = 1, 2, ..., n}, the humidity sequence is Y2 = {y2(k)|k = 1, 2, ..., n}, and the battery expansion force sequence is... The sequence is Y3={y3(k)|k=1,2,…,n}. When the battery health state is x(1), the measured at least one health factor is y1(1), y2(1) and y3(1). When the battery health state is x(2), the measured at least one health factor is y1(2), y2(2) and y3(2). And so on. When the battery health state is x(n), the measured at least one health factor is y1(n), y2(n) and y3(n).

[0050] Step 102: Based on the sequence of each health factor, determine the weight value corresponding to each health factor using the objective weighting method.

[0051] Since at least one health factor sequence is obtained under the same battery health state sequence by conducting multiple battery cycle life experiments on the target type of test battery, the weight value corresponding to each health factor can be determined by an objective weighting method based on the obtained health factor sequences. For example, for the example where the obtained at least one health factor sequence includes a temperature sequence, a humidity sequence, and a battery expansion force sequence, the weight value corresponding to the temperature health factor, the weight value corresponding to the humidity health factor, and the weight value corresponding to the battery expansion force sequence health factor can be determined by an objective weighting method based on the temperature sequence Y1={y1(k)|k=1,2,…,n}, the humidity sequence Y2={y2(k)|k=1,2,…,n}, and the battery expansion force sequence Y3={y3(k)|k=1,2,…,n}.

[0052] One possible approach is to determine the weight of each health factor using the CRITIC (Criteria Importance Through Intercrieria Correlation) weighting method, based on the sequence of health factors. The CRITIC weighting method is an objective weighting method that comprehensively measures the comparative strength and conflict between indicators. While considering the magnitude of indicator variability, the CRITIC weighting method also takes into account the correlation between indicators; it does not necessarily mean that a larger indicator value indicates greater importance, but rather relies entirely on the objective attributes of the data itself for scientific evaluation.

[0053] Step 103: Determine the initial weight values ​​of the neural network prediction model based on the weight values, and input each health factor sequence as a training sample into the neural network prediction model to obtain the corresponding predicted value.

[0054] In some embodiments, the weight values ​​corresponding to each health factor determined using the objective weighting method can be used as the initial weight values ​​of the neural network prediction model, and the sequences of each health factor can be used as training samples and input into the neural network prediction model for prediction to obtain the corresponding predicted values. The neural network prediction model can be a BP (BackPropagation) neural network model. BP neural networks have strong nonlinear mapping capabilities and can be used to simulate relationships between parameters with causal relationships, as well as relationships between parameters with autocorrelation.

[0055] Alternatively, the BP neural network model can be as follows: Figure 3 The diagram shows an input layer, hidden layers, and an output layer. Each layer has multiple nodes (neurons). Every neuron in a lower layer is fully connected to every neuron in an upper layer, as indicated by arrows. Each arrow carries a connection weight. There are no connections between neurons within the same layer. Figure 3 As shown, X1, X2, X3, ..., X n The input to the BP neural network model is represented by the value of at least one health factor corresponding to the same battery health state. The hidden layer has n neurons, where n is greater than 3. The output of the BP neural network model is only one, such as the battery health state.

[0056] In related technologies, the training process of a BP neural network model involves randomly assigning weights and thresholds to the network, inputting training samples, and repeatedly adjusting the weights and thresholds through backpropagation to make the network output as close as possible to the desired output. Training is complete when the sum of squared errors of the network's output layer is less than a specified error, and the network weights and thresholds are saved to obtain the trained BP neural network model. However, since the network weights and thresholds are randomly assigned, multiple repeated adjustments to the network weights and thresholds are required to make the network output as close as possible to the desired output. In this application, the weight values ​​corresponding to each health factor determined by the objective weighting method are directly used as the initial weight values ​​of the neural network prediction model, which can effectively reduce the repeated adjustment of the network weights and thresholds and improve the training efficiency of the neural network prediction model.

[0057] Step 104: Based on the deviation between the predicted value and each sequence value in the battery health state sequence, train the neural network prediction model to obtain the trained neural network prediction model.

[0058] In some embodiments, after obtaining the predicted value output by the neural network prediction model, the neural network prediction model can be trained based on the deviation between the predicted value and the values ​​of each sequence in the battery health state sequence, thereby obtaining a trained neural network prediction model. For example, for the example where the obtained at least one health factor sequence includes a temperature sequence, a humidity sequence, and a battery expansion force sequence, since when the battery health state is x(1), the corresponding at least one health factor is y1(1), y2(1), and y3(1), and when the battery health state is x(2), the corresponding at least one health factor is y1(2), y2(2), and y3(2), and so on, when the battery health state is x(n), the corresponding at least one health factor is y1(n), y2(n), and y3(n), thus, when y1(1), y2(1), and y3(1) are input into the neural network... After the prediction model makes a prediction and obtains the corresponding predicted value, the neural network prediction model can be trained based on the deviation between the predicted value and x(1). Similarly, after inputting y1(2), y2(2) and y3(2) into the neural network prediction model to make a prediction and obtain the corresponding predicted value, the neural network prediction model can be trained based on the deviation between the predicted value and x(2). And so on. After inputting y1(n), y2(n) and y3(n) into the neural network prediction model to make a prediction and obtain the corresponding predicted value, the neural network prediction model can be trained based on the deviation between the predicted value and x(n).

[0059] Step 105: Obtain the current health factor data of the target battery.

[0060] The target battery can be any type of battery.

[0061] The health factor data refers to the health factor sequence obtained from battery cycle life experiments. For example, assuming the health factor sequence obtained from battery cycle life experiments includes one or more of the following: temperature sequence, humidity sequence, and battery expansion force sequence, then the health factor data obtained here includes one or more of the following: temperature data, humidity data, and battery expansion force data.

[0062] In some embodiments, health factor data of the target battery at the current moment can be obtained to predict the battery health status of the target battery.

[0063] Step 106: Input the health factor data into the trained neural network prediction model to predict the battery health status of the target battery.

[0064] In some embodiments, the acquired health factor data of the target battery can be input into a trained neural network prediction model to predict the battery health status. For example, since the neural network prediction model is trained using health factor sequences obtained from battery cycle life experiments as training samples to predict battery health status, the acquired health factor data of the target battery at the current moment, corresponding to the health factor sequences obtained from battery cycle life experiments, can be input into the trained neural network prediction model to predict the battery health status of the target battery.

[0065] In this embodiment, by conducting multiple battery cycle life experiments on a test battery of the target type, at least one health factor sequence is obtained for the test battery under the same battery health state sequence. Based on each health factor sequence, an objective weighting method is used to determine the weight value corresponding to each health factor. Then, based on the weight values, the initial weight values ​​of the neural network prediction model are determined. Each health factor sequence is used as a training sample and input into the neural network prediction model to obtain the corresponding predicted value. Furthermore, based on the deviation between the predicted value and the values ​​of each sequence in the battery health state sequence, the neural network prediction model is trained to obtain a trained neural network prediction model. After acquiring the current health factor data of the target battery, the health factor data is input into the trained neural network prediction model to predict the battery health state of the target battery. Therefore, accurate monitoring of battery health state can be achieved. At the same time, since the initial weight values ​​of the neural network prediction model can be determined based on the weight values ​​corresponding to each health factor using the objective weighting method, the training efficiency of the neural network prediction model can be effectively improved.

[0066] The above analysis shows that in this invention, the weight value of each health factor can be determined using an objective weighting method based on the health factor sequences. The following section combines... Figure 4 This process will be explained.

[0067] Figure 4 This is a flowchart illustrating another battery health status prediction method provided in an embodiment of the present invention. It should be noted that the objective weighting method used in this embodiment is the CRITIC weighting method. The CRITIC weighting method is a method that comprehensively measures the objective weight of evaluation indicators based on the comparative strength and conflict between indicators. The comparative strength refers to the magnitude of the difference between the various values ​​of the same indicator, expressed in the form of standard deviation. The larger the standard deviation, the greater the fluctuation, meaning the greater the difference between the values ​​of the indicators, and the higher the weight. The conflict between indicators is represented by the correlation coefficient. If two indicators have a strong positive correlation, it indicates a smaller conflict, and the lower the weight.

[0068] For the CRITIC weighting method, with a fixed standard deviation, the lower the conflict between indicators, the lower the weight; conversely, the higher the conflict, the higher the weight. Furthermore, the stronger the positive correlation between two indicators (the closer the correlation coefficient is to 1), the lower the conflict, indicating that the information reflected by these two indicators is highly similar.

[0069] like Figure 4 As shown, the battery health status prediction method may include the following steps:

[0070] Step 401: Perform multiple battery cycle life experiments on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence.

[0071] It should be noted that the execution process of this step can be referred to step 101 in the previous embodiment, as the principle is the same, and will not be repeated here.

[0072] Step 402: Normalize each health factor sequence to obtain the corresponding target health factor sequence.

[0073] To eliminate the influence of different dimensions on the prediction results, in some embodiments, it is necessary to perform dimensionless processing on each health factor sequence. Optionally, each health factor sequence can be positively normalized or negatively normalized to obtain the corresponding target health factor sequence.

[0074] As an example, assuming that the obtained test battery has at least one health factor sequence under the same battery health state sequence, including temperature sequence Y1={y1(k)|k=1,2,…,n}, humidity sequence Y2={y2(k)|k=1,2,…,n}, and battery expansion force sequence Y3={y3(k)|k=1,2,…,n}, then the following data matrix X can be constructed:

[0075]

[0076] Where, x 11 x 12 ... x 1n Let Y1 = {y1(k) | k = 1, 2, ..., n}, x 11 That is, y1(1), x 12 That is, y1(2), and so on, x 1n That is, y1(n); x 21 x 22 ... x 2n Let the humidity sequence be Y2={y2(k)|k=1,2,…,n}, x 21 That is, y2(1), x 22 That is, y2(2), and so on, x 2n That is, y2(n); x 31 x 32 ... x 3n The battery expansion force sequence is represented by Y3 = {y3(k) | k = 1, 2, ..., n}, x 31 That is, y3(1), x 32 That is, y3(2), and so on, x 3n That is, y3(n).

[0077] This allows for the normalization of each element in the data matrix X. Optionally, if x ij For a positive indicator, meaning the larger the value, the better, the following formula can be used to perform negative normalization on the indicator:

[0078]

[0079] Where, x ij ′ represents the element x in the data matrix. ij The data obtained after normalization, x ij Let x represent the data in the i-th row and j-th column of the data matrix X, min(x) j ) represents the minimum value of the j-th column in the data matrix X, max(x) j () represents the maximum value of the data in the j-th column of the data matrix X.

[0080] And / or,

[0081] If x ij For a negative indicator, meaning the smaller the value, the better, the indicator can be negatively normalized using the following formula:

[0082]

[0083] Where, x ij ′ represents the element x in the data matrix. ij The data obtained after normalization, x ij Let x represent the data in the i-th row and j-th column of the data matrix X, min(x) j ) represents the minimum value of the j-th column in the data matrix X, max(x) j () represents the maximum value of the data in the j-th column of the data matrix X.

[0084] This results in the following normalized data matrix X′:

[0085]

[0086] Where, x 11 ′、x 12 ′、…、x 1n ′ represents the target health factor sequence corresponding to the temperature sequence, x 21 ′、x 22 ′、…、x 2n ′ represents the target health factor sequence corresponding to the humidity sequence, x 31 ′、x 32 ′、…、x 3n ′ represents the target health factor sequence corresponding to the battery expansion force sequence.

[0087] Step 403: Use the standard deviation of any target health factor sequence as the indicator difference of the corresponding health factor.

[0088] Among them, the indicator variability is used to indicate the variation and fluctuation of the internal values ​​of any health factor.

[0089] In the CRITIC weighting method, the standard deviation is used to represent the variation and fluctuation of the values ​​of each indicator. The larger the standard deviation, the greater the numerical difference of the indicator, the more information it can reflect, and the stronger the evaluation intensity of the indicator itself. Therefore, more weight should be assigned to the indicator. Thus, in some embodiments, the standard deviation of any target health factor sequence can be used as the indicator difference of the corresponding health factor.

[0090] Alternatively, the standard deviation of any target health factor sequence can be calculated using the following formula:

[0091]

[0092] in, Represents the target health factor sequence x j The average value, x ij ′ represents the target health factor sequence x j One element value in S j Represents the target health factor sequence x j The standard deviation, where n represents the target health factor sequence x. j The number of elements.

[0093] Step 404: Determine the index conflict of the corresponding health factor based on the correlation coefficient between any target health factor sequence and each target health factor sequence other than the target health factor sequence.

[0094] The correlation coefficient is used to indicate the degree of correlation between any target health factor sequence and other target health factor sequences besides the target health factor sequence.

[0095] In the CRITIC weighting method, correlation coefficients are used to represent the correlation between indicators. The stronger the correlation with other indicators, the less conflict there is between the indicator and other indicators, the more identical information it reflects, and the more repetitive the evaluation content it can reflect. To a certain extent, this weakens the evaluation strength of the indicator, and the weight assigned to the indicator should be reduced. Thus, in some embodiments, the indicator conflict of the corresponding health factor can be determined based on the correlation coefficient between any target health factor sequence and each target health factor sequence other than the target health factor sequence.

[0096] As one possible approach, the correlation coefficient between any target health factor sequence and all other target health factor sequences can be determined using the following formula:

[0097]

[0098] Where, r ij Represents the target health factor sequence x j With target health factor sequence x i The correlation coefficient between them Represents the target health factor sequence x j The average value, Represents the target health factor sequence x i The average value, where n represents the number of elements in the target health factor sequence (each target health factor sequence has the same number of elements).

[0099] Therefore, based on the correlation coefficients between the target health factor sequence and all other target health factor sequences, the conflict value between the target health factor sequence and all other target health factor sequences can be determined. Then, the conflict values ​​between the target health factor sequence and all other target health factor sequences can be summed to obtain the index conflict of the health factor corresponding to the target health factor sequence. This process can be expressed by the following formula:

[0100]

[0101] Among them, R j Represents the target health factor sequence x j The conflicting indicators of the corresponding health factors, r ij Represents the target health factor sequence x j With target health factor sequence x i The correlation coefficient between them, where m represents the sequence excluding the target health factor sequence x. j The number of each target health factor sequence other than the target health factor sequence.

[0102] Step 405: For any health factor, determine the information content of the health factor based on the differences and conflicts between the indicators of the health factor.

[0103] Information content is used to indicate the importance of any health factor. In the CRITIC weighting method, the greater the information content, the greater the role of the corresponding indicator in the entire indicator system, and the more weight should be assigned to it.

[0104] As one possible approach, for any health factor, the information content of the health factor can be determined by multiplying the variability and conflict of its indicators. This process can be expressed by the following formula:

[0105]

[0106] Among them, C j Represents the target health factor sequence x j The information content of the corresponding health factors, S j Represents the target health factor sequence x j standard deviation, r ij Represents the target health factor sequence x j With target health factor sequence x i The correlation coefficient between them, where m represents the sequence excluding the target health factor sequence x. j The number of each target health factor sequence other than R j Represents the target health factor sequence x j The corresponding health factor indicators are conflicting.

[0107] Step 406: Determine the weight value corresponding to each health factor based on the information content of each health factor.

[0108] One possible approach is to sum the information content of each health factor to obtain the target information content. Then, for any given health factor, the ratio between the information content of that health factor and the target information content is determined as the weight value for that health factor. This process can be expressed by the following formula:

[0109]

[0110] Among them, W j Represents the target health factor sequence x j The corresponding weight value of the health factor, C j Represents the target health factor sequence x j The information content of the corresponding health factors, where p represents the number of health factors corresponding to the target health factor sequence.

[0111] Step 407: Determine the initial weight values ​​of the neural network prediction model based on the weight values, and input each health factor sequence as a training sample into the neural network prediction model to obtain the corresponding predicted value.

[0112] Step 408: Based on the deviation between the predicted value and each sequence value in the battery health state sequence, train the neural network prediction model to obtain a trained neural network prediction model.

[0113] Step 409: Obtain the current health factor data of the target battery.

[0114] Step 410: Input the health factor data into the trained neural network prediction model to predict the battery health status of the target battery.

[0115] It should be noted that the execution process of steps 407-410 can be referred to steps 103-106 in the previous embodiment, as the principle is the same, and will not be repeated here.

[0116] In this embodiment, by normalizing each health factor sequence, a corresponding target health factor sequence is obtained. The standard deviation of any target health factor sequence is used as the indicator difference of the corresponding health factor. Furthermore, based on the correlation coefficient between any target health factor sequence and all other target health factor sequences, the indicator conflict of the corresponding health factor is determined. Thus, for any given health factor, the information content of the health factor is determined based on its indicator difference and indicator conflict. Finally, based on the information content of each health factor, the corresponding weight value is determined. Therefore, the CRITIC weighting method can be used to determine the weight value corresponding to each health factor, facilitating the determination of the initial weight values ​​for the neural network prediction model and improving the training efficiency of the neural network prediction model.

[0117] To clearly illustrate how multiple battery cycle life experiments are conducted and how a trained neural network prediction model is obtained in this invention, this embodiment provides another method for predicting battery health status. Figure 5 This is a flowchart illustrating another battery health status prediction method provided in an embodiment of the present invention.

[0118] like Figure 5 As shown, the battery health status prediction method may include the following steps:

[0119] Step 501: Based on preset experimental control factors, conduct multiple battery cycle life experiments on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence.

[0120] The preset experimental control factors include one or more of the following: experimental method, charging method, discharging method, temperature, humidity, battery expansion force, charging cut-off voltage, and discharging cut-off voltage.

[0121] Among them, at least one health factor sequence includes one or more of temperature sequence, humidity sequence, and battery expansion force sequence. In this case, the corresponding health factor is as follows: Figure 2 As shown, this includes one or more of the following: ambient temperature, ambient humidity, and battery expansion force.

[0122] As an example, to analyze the effects of ambient temperature, ambient humidity, and battery expansion force on battery health status, multiple battery cycle life experiments can be conducted on the target type of test battery based on the preset experimental control factors shown in Table 1 to obtain at least one health factor sequence of the test battery under the same battery health status sequence.

[0123] Table 1. Values ​​of different factors for battery cycle life testing.

[0124] Way Full charge and discharge, then let stand. Charging method Constant current-constant voltage Discharge method constant current temperature 5℃、15℃、25℃、35℃ humidity 40％、60％、80％、95％ Battery expansion force (output) Actual expansion force of the battery Charging cutoff voltage Cell charging cutoff voltage Discharge cutoff voltage Cell discharge cutoff voltage

[0125] Further explanations of this step can be found in the description of step 101 in the above embodiments, and will not be repeated here.

[0126] Step 502: Based on the sequence of each health factor, determine the weight value corresponding to each health factor using the objective weighting method.

[0127] Step 503: Determine the initial weight values ​​of the neural network prediction model based on the weight values, and input each health factor sequence as a training sample into the neural network prediction model to obtain the corresponding predicted value.

[0128] It should be noted that the execution process of steps 502-503 can be referred to steps 102-103 in the above embodiment, as the principle is the same, and will not be repeated here.

[0129] Step 504: Determine whether the deviation between the predicted value corresponding to each sequence value in each health factor sequence and each sequence value in the battery health status sequence reaches the preset expected value.

[0130] In some embodiments, after obtaining the predicted values ​​output by the neural network prediction model, it can be determined whether the deviation between the predicted values ​​corresponding to each sequence value in each health factor sequence and each sequence value in the battery health state sequence reaches a preset expected value. The present invention does not limit the setting of the preset expected value. Optionally, the preset expected value can be set based on human experience; for example, it can be set to 1%. Alternatively, the preset expected value can be dynamically adjusted according to actual application needs; this embodiment does not limit this.

[0131] As an example, suppose the battery health state sequence is X = {x(k)|k = 1, 2, ..., n}, and the measured health factor sequence includes a temperature sequence, a humidity sequence, and a battery expansion force sequence, where the temperature sequence is Y1 = {y1(k)|k = 1, 2, ..., n}, the humidity sequence is Y2 = {y2(k)|k = 1, 2, ..., n}, and the battery expansion force sequence is Y3 = {y3(k)|k = 1, 2, ..., n}. That is, when the battery health state is x(1), the measured health factors are y1(1), y2(1), and y3(1); when the battery health state is x(2), the measured health factors are y1(2), y2(2), and y3(2), and so on, when the battery health state is x(n)... When measuring at least one health factor, y1(n), y2(n), and y3(n), y1(n), y2(n), and y3(n) are input into the BP neural network model for prediction. After obtaining the corresponding predicted value, it can be determined whether the deviation between the obtained predicted value and x(n) reaches the preset expected value. Similarly, y1(2), y2(2), and y3(2) are input into the BP neural network model for prediction. After obtaining the corresponding predicted value, it can be determined whether the deviation between the obtained predicted value and x(n) reaches the preset expected value. And so on, y1(n), y2(n), and y3(n) are input into the BP neural network model for prediction. After obtaining the corresponding predicted value, it can be determined whether the deviation between the obtained predicted value and x(n) reaches the preset expected value.

[0132] Step 505: If the deviation does not reach the preset expected value, the weight values ​​and thresholds of the neural network prediction model are corrected according to the deviation, and the health factor sequences are re-inputted into the neural network prediction model to determine the new deviation.

[0133] In some embodiments, if the deviation does not reach the preset expected value, the weights and thresholds of the BP neural network model are corrected based on the deviation, and the health factor sequences are re-inputted into the neural network prediction model to determine the new deviation. As one possible implementation, the specific method for correcting the weights and thresholds is as follows:

[0134] W i+1 =W i -(H+λdiag[H]) -1 d

[0135] Where H represents the Hessian matrix, diag[H] represents the diagonal matrix of the Hessian matrix, and λ represents the variable parameter that changes according to the training error, ranging from

[10] . -6 10 6 ], where d represents the descent gradient of the modified parameter.

[0136] Step 506: When the deviation reaches the preset expected value, complete the training of the neural network prediction model to obtain the trained neural network prediction model.

[0137] In some embodiments, if the deviation reaches the preset expected value, it indicates that the training of the BP neural network model has been completed, and a trained BP neural network model can be obtained, that is, a trained neural network prediction model.

[0138] Step 507: Obtain the current health factor data of the target battery.

[0139] Step 508: Input the health factor data into the trained neural network prediction model to predict the battery health status of the target battery.

[0140] It should be noted that the execution process of steps 507-508 can be referred to steps 105-106 in the above embodiment, as the principle is the same, and will not be repeated here.

[0141] In this embodiment, multiple battery cycle life experiments are conducted on the target type of test battery based on preset experimental control factors to obtain at least one health factor sequence under the same battery health state sequence, enabling data acquisition based on experiments. Furthermore, by determining the initial weight values ​​of the neural network prediction model according to the weight values, each health factor sequence is used as a training sample and input into the neural network prediction model for prediction to obtain the corresponding predicted value. The system then determines whether the deviation between the predicted value corresponding to each sequence value in each health factor sequence and the sequence values ​​in the battery health state sequence reaches a preset expected value. If the deviation does not reach the preset expected value, the weight values ​​and thresholds of the neural network prediction model are corrected based on the deviation, and each health factor sequence is re-inputted into the neural network prediction model to determine a new deviation. Alternatively, if the deviation reaches the preset expected value, the training of the neural network prediction model is completed, resulting in a trained neural network prediction model. Thus, the training process of the neural network prediction model is completed.

[0142] To achieve the above embodiments, the present invention also proposes a battery health status prediction device.

[0143] Figure 6 This is a schematic diagram of a battery health status prediction device provided in an embodiment of the present invention.

[0144] like Figure 6 As shown, the battery health status prediction device includes: a first processing module 61, a determination module 62, a second processing module 63, a training module 64, an acquisition module 65, and a prediction module 66.

[0145] The first processing module 61 is used to perform multiple battery cycle life experiments on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence.

[0146] The determination module 62 is used to determine the weight value of each health factor based on the sequence of each health factor using an objective weighting method.

[0147] The second processing module 63 is used to determine the initial weight values ​​of the neural network prediction model based on the weight values, and to input each health factor sequence as a training sample into the neural network prediction model for prediction in order to obtain the corresponding predicted value.

[0148] Training module 64 is used to train the neural network prediction model based on the deviation between the predicted value and each sequence value in the battery health state sequence, so as to obtain a trained neural network prediction model.

[0149] The acquisition module 65 is used to acquire the health factor data of the target battery at the current moment, wherein the target battery is a battery of the target type;

[0150] The prediction module 66 is used to input health factor data into a trained neural network prediction model to predict the battery health status of the target battery.

[0151] Furthermore, in one possible implementation of this invention, the objective weighting method is the CRITIC weighting method; the determining module 62 includes:

[0152] The first processing unit is used to normalize each health factor sequence to obtain the corresponding target health factor sequence.

[0153] The second processing unit is used to take the standard deviation of any target health factor sequence as the index difference of the corresponding health factor, wherein the index difference is used to indicate the difference fluctuation of the internal values ​​of any health factor.

[0154] The first determining unit is used to determine the index conflict of the corresponding health factor based on the correlation coefficient between any target health factor sequence and each target health factor sequence other than the target health factor sequence, wherein the correlation coefficient is used to indicate the degree of correlation between any target health factor sequence and each target health factor sequence other than the target health factor sequence.

[0155] The second determining unit is used to determine the information content of any health factor based on the differences and conflicts between the indicators of the health factors. The information content is used to indicate the importance of any health factor.

[0156] The third determining unit is used to determine the weight value corresponding to each health factor based on the information content of each health factor.

[0157] Furthermore, in one possible implementation of this invention, the first determining unit is further configured to:

[0158] For any given target health factor sequence, the correlation coefficient between the target health factor sequence and all other target health factor sequences is determined using the following formula:

[0159]

[0160] Where, r ij Represents the target health factor sequence x j With target health factor sequence x i The correlation coefficient between them Represents the target health factor sequence x j The average value, Represents the target health factor sequence x i The average value, where n represents the number of elements in the target health factor sequence;

[0161] Based on the correlation coefficient between the target health factor sequence and each other target health factor sequence, the conflict value between the target health factor sequence and each other target health factor sequence is determined.

[0162] The conflict values ​​between the target health factor sequence and each other target health factor sequence are summed to obtain the index conflict of the health factor corresponding to the target health factor sequence.

[0163] Furthermore, in one possible implementation of this invention, the second determining unit is further configured to:

[0164] For any health factor, the product of the indicator variability and indicator conflict of the health factor is determined as the information content of the health factor.

[0165] Furthermore, in one possible implementation of this invention, the third determining unit is further configured to:

[0166] The target information content is obtained by summing the information content of each health factor;

[0167] For any health factor, the ratio between the information content of the health factor and the target information content is determined as the weight value corresponding to the health factor.

[0168] Furthermore, in one possible implementation of this invention, the neural network prediction model is a BP neural network model, and the training module 64 is further used for:

[0169] Determine whether the deviation between the predicted value corresponding to each sequence value in each health factor sequence and each sequence value in the battery health status sequence reaches the preset expected value;

[0170] If the deviation does not reach the preset expected value, the weight values ​​and thresholds of the BP neural network model are corrected according to the deviation, and the health factor sequences are re-inputted into the neural network prediction model to determine the new deviation.

[0171] If the deviation reaches the preset expected value, the training of the BP neural network model is completed, and a trained BP neural network model is obtained.

[0172] Furthermore, in one possible implementation of this invention, at least one health factor sequence includes one or more of a temperature sequence, a humidity sequence, and a battery expansion force sequence; the first processing module 61 is further configured to:

[0173] Based on preset experimental control factors, multiple battery cycle life experiments are conducted on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence. The preset experimental control factors include one or more of the experimental method, charging method, discharging method, temperature, humidity, battery expansion force, charging cut-off voltage and discharging cut-off voltage.

[0174] It should be noted that the foregoing explanation of the battery health status prediction method embodiment also applies to the battery health status prediction device of this embodiment, and will not be repeated here.

[0175] In this embodiment, by conducting multiple battery cycle life experiments on a test battery of the target type, at least one health factor sequence is obtained for the test battery under the same battery health state sequence. Based on each health factor sequence, an objective weighting method is used to determine the weight value corresponding to each health factor. Then, based on the weight values, the initial weight values ​​of the neural network prediction model are determined. Each health factor sequence is used as a training sample and input into the neural network prediction model to obtain the corresponding predicted value. Furthermore, based on the deviation between the predicted value and the values ​​of each sequence in the battery health state sequence, the neural network prediction model is trained to obtain a trained neural network prediction model. After acquiring the current health factor data of the target battery, the health factor data is input into the trained neural network prediction model to predict the battery health state of the target battery. Therefore, accurate monitoring of battery health state can be achieved. At the same time, since the initial weight values ​​of the neural network prediction model can be determined based on the weight values ​​corresponding to each health factor using the objective weighting method, the training efficiency of the neural network prediction model can be effectively improved.

[0176] To implement the above embodiments, the present invention also proposes an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute the battery health state prediction method proposed in any of the above embodiments of the present invention.

[0177] To implement the above embodiments, the present invention also proposes a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the battery health state prediction method proposed in any of the above embodiments of the present invention.

[0178] To implement the above embodiments, the present invention also proposes a computer program product, including a computer program that, when executed by a processor, implements the battery health state prediction method proposed in any of the above embodiments of the present invention.

[0179] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0180] It should be noted that, Figure 7 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0181] like Figure 7 As shown, the electronic device includes:

[0182] The memory 701, the processor 702, and the computer program stored on the memory 701 and executable on the processor 702.

[0183] When the processor 702 executes the program, it implements the battery health status prediction method provided in any of the above embodiments.

[0184] Furthermore, electronic devices also include:

[0185] Communication interface 703 is used for communication between memory 701 and processor 702.

[0186] The memory 701 is used to store computer programs that can run on the processor 702.

[0187] The memory 701 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0188] The processor 702 is used to implement the battery health state prediction method described in any of the above embodiments when executing the program.

[0189] If the memory 701, processor 702, and communication interface 703 are implemented independently, then the communication interface 703, memory 701, and processor 702 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0190] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0191] The processor 702 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0192] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0193] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0194] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0195] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0196] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0197] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0198] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0199] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for predicting battery health status, characterized in that, include: Multiple battery cycle life experiments were conducted on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence, wherein the at least one health factor sequence includes one or more of temperature sequence, humidity sequence, and battery expansion force sequence; Based on the health factor sequences, the CRITIC objective weighting method is used to determine the weight value corresponding to each health factor, including: normalizing each health factor sequence to obtain the corresponding target health factor sequence; using the standard deviation of any target health factor sequence as the indicator variability of the corresponding health factor, wherein the indicator variability is used to indicate the variation and fluctuation of the internal values ​​of any health factor; determining the indicator conflict of the corresponding health factor based on the correlation coefficient between any target health factor sequence and other target health factor sequences, wherein the correlation coefficient is used to indicate the degree of correlation between any target health factor sequence and other target health factor sequences; for any health factor, determining the information content of the health factor based on the indicator variability and indicator conflict, wherein the information content is used to indicate the importance of any health factor; and determining the weight value corresponding to each health factor based on the information content of each health factor. Based on the weight values, the initial weight values ​​of the neural network prediction model are determined, and each of the health factor sequences is used as a training sample and input into the neural network prediction model for prediction to obtain the corresponding predicted value. The neural network prediction model is a BP neural network model. Determine whether the deviation between the predicted value corresponding to each sequence value in each of the health factor sequences and each sequence value in the battery health status sequence reaches a preset expected value; If the deviation does not reach the preset expected value, the weight values ​​and thresholds of the neural network prediction model are corrected according to the deviation, and each of the health factor sequences is re-inputted into the neural network prediction model to determine a new deviation. When the deviation reaches the preset expected value, the training of the neural network prediction model is completed, and a trained neural network prediction model is obtained. Obtain the current health factor data of the target battery, wherein the target battery is a battery of the target type; The health factor data is input into the trained neural network prediction model to predict the battery health status of the target battery.

2. The method according to claim 1, characterized in that, The step of determining the index conflict of a corresponding health factor based on the correlation coefficient between any of the target health factor sequences and all other target health factor sequences includes: For any of the target health factor sequences, the correlation coefficient between the target health factor sequence and each of the other target health factor sequences is determined using the following formula: in, Represents the target health factor sequence With target health factor sequence The correlation coefficient between them Represents the target health factor sequence The average value, Represents the target health factor sequence The average value, where n represents the number of elements in the target health factor sequence; Based on the correlation coefficient between the target health factor sequence and each other target health factor sequence, the conflict value between the target health factor sequence and each other target health factor sequence is determined. The conflict values ​​between the target health factor sequence and each other target health factor sequence are added together to obtain the index conflict of the health factor corresponding to the target health factor sequence.

3. The method according to claim 1, characterized in that, For any given health factor, determining the information content of the health factor based on the differences and conflicts in its indicators includes: For any health factor, the product of the indicator variability and indicator conflict of the health factor is determined as the information content of the health factor.

4. The method according to claim 1, characterized in that, The step of determining the weight value corresponding to each health factor based on the information content of each health factor includes: The target information content is obtained by summing the information content of each health factor; For any health factor, the ratio between the information content of the health factor and the target information content is determined as the weight value corresponding to the health factor.

5. The method according to any one of claims 1-4, characterized in that, The process of performing multiple battery cycle life experiments on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence includes: Based on preset experimental control factors, multiple battery cycle life experiments are conducted on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence. The preset experimental control factors include one or more of the following: experimental method, charging method, discharging method, temperature, humidity, battery expansion force, charging cut-off voltage, and discharging cut-off voltage.

6. A battery health status prediction device, characterized in that, include: The first processing module is used to perform multiple battery cycle life experiments on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence, wherein the at least one health factor sequence includes one or more of temperature sequence, humidity sequence, and battery expansion force sequence. The determination module is used to determine the weight value corresponding to each health factor based on the health factor sequence using an objective weighting method; The second processing module is used to determine the initial weight values ​​of the neural network prediction model based on the weight values, and to input each of the health factor sequences as training samples into the neural network prediction model for prediction to obtain the corresponding predicted values. The neural network prediction model is a BP neural network model. The training module is used to determine whether the deviation between the predicted value corresponding to each sequence value in each of the health factor sequences and each sequence value in the battery health status sequence reaches a preset expected value; if the deviation does not reach the preset expected value, the weight values ​​and thresholds of the neural network prediction model are corrected according to the deviation, and each of the health factor sequences is re-inputted into the neural network prediction model to determine a new deviation; if the deviation reaches the preset expected value, the training of the neural network prediction model is completed to obtain a trained neural network prediction model. The acquisition module is used to acquire the health factor data of the target battery at the current moment, wherein the target battery is a battery of the target type; The prediction module is used to input the health factor data into the trained neural network prediction model to predict the battery health status of the target battery. The objective weighting method is the CRITIC weighting method; the determining module includes: The first processing unit is used to normalize each of the health factor sequences to obtain the corresponding target health factor sequences. The second processing unit is used to take the standard deviation of any of the target health factor sequences as the index difference of the corresponding health factor, wherein the index difference is used to indicate the difference fluctuation of the internal values ​​of any health factor. The first determining unit is configured to determine the index conflict of a corresponding health factor based on the correlation coefficient between any target health factor sequence and each target health factor sequence other than the target health factor sequence, wherein the correlation coefficient is used to indicate the degree of correlation between any target health factor sequence and each target health factor sequence other than the target health factor sequence. The second determining unit is used to determine the information content of any health factor based on the index differences and index conflicts of the health factor, wherein the information content is used to indicate the importance of any health factor. The third determining unit is used to determine the weight value corresponding to each health factor based on the information content of each health factor.

7. The apparatus according to claim 6, characterized in that, The first determining unit is further configured to: For any of the target health factor sequences, the correlation coefficient between the target health factor sequence and each of the other target health factor sequences is determined using the following formula: in, Represents the target health factor sequence With target health factor sequence The correlation coefficient between them Represents the target health factor sequence The average value, Represents the target health factor sequence The average value, where n represents the number of elements in the target health factor sequence; Based on the correlation coefficient between the target health factor sequence and each other target health factor sequence, the conflict value between the target health factor sequence and each other target health factor sequence is determined. The conflict values ​​between the target health factor sequence and each other target health factor sequence are added together to obtain the index conflict of the health factor corresponding to the target health factor sequence.

8. The apparatus according to claim 6, characterized in that, The second determining unit is further configured to: For any health factor, the product of the indicator variability and indicator conflict of the health factor is determined as the information content of the health factor.

9. The apparatus according to claim 6, characterized in that, The third determining unit is further configured to: The target information content is obtained by summing the information content of each health factor; For any health factor, the ratio between the information content of the health factor and the target information content is determined as the weight value corresponding to the health factor.

10. The apparatus according to any one of claims 6-9, characterized in that, The first processing module is further configured to: Based on preset experimental control factors, multiple battery cycle life experiments are conducted on the target type of test battery to obtain at least one health factor sequence of the test battery under the same battery health state sequence. The preset experimental control factors include one or more of the following: experimental method, charging method, discharging method, temperature, humidity, battery expansion force, charging cut-off voltage, and discharging cut-off voltage.

11. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.

12. A computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-5.

13. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1-5.

Images