Battery SOH estimation method, control device and storage medium

Abstract

The invention relates to the technical field of vehicle batteries, particularly provides a battery SOH estimation method, a control device and a storage medium, and aims to solve the problem that in the prior art, the use scene for detecting the SOH of a vehicle battery is limited. In order to achieve the purpose, the method comprises the steps that in the charging process, ICA data of a battery are obtained; in response to a supplement instruction, supplementing the ICA data to obtain supplement data; and obtaining the SOH estimation value of the battery according to the supplementary data. According to the method, the complete ICA curve can be complemented and obtained by only measuring a part of the values of the preset charging time, the accurate SOH value is finally obtained, the battery does not need to be fully charged, the application scene of SOH prediction is expanded, and a user is further away from mileage anxiety.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle batteries, and specifically provides a method for estimating the state of health (SOH) of a battery, a control device, and a storage medium. Background Art

[0002] With the vigorous development and popularization of electric vehicles, users have higher and higher requirements for the safety performance of batteries. The power batteries currently used in electric vehicles are mainly divided into lithium iron phosphate system batteries (LFP) and ternary system batteries (NCM). Since LFP has more excellent performance in terms of safety, even though its energy density is average, more and more electric vehicles using LFP have emerged in recent years.

[0003] The estimation of the state of health (SOH) of vehicle batteries is a core issue in the battery management system. Accurately estimating the battery SOH is beneficial for the vehicle to formulate appropriate control strategies, improve the accuracy of SOC calculation, extend the driving range, and reduce the user's range anxiety. Existing methods for calculating SOH require the battery to be fully charged to obtain a complete ICA curve (the discharge curve of the battery), and then estimate the SOH. That is, if the user disconnects from charging without fully charging, the vehicle will not be able to accurately calculate the value of SOH at this time. Even in most cases, the vehicle usually disconnects the power after being fully charged, there are still many scenarios where the user cannot wait too long and terminates charging midway, unable to reach the full charge state, which limits the application scenarios of the vehicle in evaluating SOH.

[0004] The above limitations make it impossible to accurately estimate the SOH of LFP batteries when they are not fully charged, and the problem of estimating the SOH of LFP batteries has always been a difficult problem in the industry.

[0005] Correspondingly, a new solution is needed in this field to solve the above problems. Summary of the Invention

[0006] The present application aims to solve or partially solve the above technical problems, that is, the problem that the SOH of LFP batteries cannot be accurately estimated when they are not fully charged.

[0007] In a first aspect, the present invention provides a method for estimating the SOH of a battery, the method comprising:

[0008] During the charging process, obtaining ICA data of the battery;

[0009] In response to a supplementary instruction, supplementing the ICA data to obtain supplementary data;

[0010] According to the supplementary data, obtaining an estimated value of the SOH of the battery.

[0011] In the preferred technical solution of the above battery SOH estimation method, supplementing the ICA data includes:

[0012] Supplementing the ICA data through an ICA data acquisition model.

[0013] In the preferred technical solution of the above battery SOH estimation method, the ICA data includes a sequence of the first number of data within the first voltage range. When the first number is less than the preset number, the supplement instruction is triggered.

[0014] In the preferred technical solution of the above battery SOH estimation method, the first data in the ICA data is the dQ / dV value corresponding to the lower limit voltage of the first voltage range, and the ICA data is the first number of dQ / dV values. When the first number is greater than or equal to the second number, supplementing the ICA data through an ICA data acquisition model includes:

[0015] Step A, obtaining the previous second number of data in the ICA data as the current data;

[0016] Step B, inputting the current data into the ICA data acquisition model to obtain the next data of the current data;

[0017] Step C, using the next data as the last data of the second number of data, and re-obtaining the second number of data as the current data;

[0018] Repeat Step B and Step C until the dQ / dV value corresponding to the upper limit voltage of the first voltage range is obtained.

[0019] In the preferred technical solution of the above battery SOH estimation method, obtaining the historical dQ / dV data sequence corresponding to the third voltage range obtained according to the actual sampling frequency during the battery charging process;

[0020] Based on the historical dQ / dV data sequence, obtaining a sample dQ / dV data sequence through interpolation operation according to the first voltage range and the voltage sampling step size;

[0021] Training the ICA data acquisition model and the battery SOH estimation model respectively based on the sample dQ / dV data sequence;

[0022] Wherein, the lower limit voltage of the third voltage range is less than or equal to the lower limit voltage of the first voltage range, and the upper limit voltage of the third voltage range is greater than or equal to the upper limit voltage of the first voltage range.

[0023] In the preferred technical solution of the above battery SOH estimation method, obtaining the SOH estimation value of the battery includes: obtaining the SOH estimation value of the battery through a convolutional neural network.

[0024] In the preferred technical solution of the above battery SOH estimation method, the convolutional neural network includes a LeNet network. The LeNet network sequentially includes a plurality of convolutional blocks and fully connected layers. The number of input channels of the first convolutional block among the plurality of convolutional blocks is denoted as the first channel number, and the number of output channels of the last convolutional block among the plurality of convolutional blocks is denoted as the second channel number. The first channel number is less than the second channel number.

[0025] In the preferred technical solution of the above battery SOH estimation method, the method further includes a model training method, and the model training method includes:

[0026] Obtaining historical ICA data corresponding to a third voltage interval obtained according to an actual sampling frequency during the charging process of the battery;

[0027] Based on the historical ICA data, performing interpolation operation on the historical ICA data to obtain sample ICA data;

[0028] Training the ICA data acquisition model based on the sample ICA data;

[0029] Wherein, the lower limit voltage of the third voltage interval is less than or equal to the lower limit voltage of the first voltage interval, and the upper limit voltage of the third voltage interval is greater than or equal to the upper limit voltage of the first voltage interval.

[0030] In the preferred technical solution of the above battery SOH estimation method, the upper limit voltage of the first voltage interval is the charging cut-off voltage of the battery.

[0031] In a second aspect, the present application proposes a control device, including at least one processor and at least one storage device. The storage device is adapted to store multiple program codes, and the program codes are adapted to be loaded and run by the processor to execute the battery SOH estimation method described in any one of the above technical solutions.

[0032] In a third aspect, the present application proposes a storage medium, which is adapted to store multiple program codes, and the program codes are adapted to be loaded and run by a processor to execute the battery SOH estimation method described in any one of the above solutions.

[0033] The present application can finally complete and obtain a complete ICA curve and ultimately obtain an accurate SOH value by only measuring the values of a part of the preset charging time, without waiting for the battery to be fully charged, expanding the application scenario of SOH prediction and enabling users to further stay away from range anxiety.

[0034] Solution 1. A method for estimating the SOH of a battery, characterized in that the method comprises:

[0035] During the charging process, obtain the ICA data of the battery;

[0036] In response to a supplement instruction, supplement the ICA data to obtain supplementary data;

[0037] According to the supplementary data, obtain the SOH estimated value of the battery.

[0038] Solution 2. The method for estimating the SOH of a battery according to Solution 1, characterized in that supplementing the ICA data comprises:

[0039] Supplement the ICA data through an ICA data acquisition model.

[0040] Solution 3. The method for estimating the SOH of a battery according to Solution 1 or 2, characterized in that the ICA data includes a sequence of the first number of data within a first voltage range, and when the first number is less than a preset number, trigger the supplement instruction.

[0041] Solution 4. The method for estimating the SOH of a battery according to Solution 3, characterized in that the first data in the ICA data is the dQ / dV value corresponding to the lower limit voltage of the first voltage range, and the ICA data is the first number of dQ / dV values. When the first number is greater than or equal to the second number, supplement the ICA data through an ICA data acquisition model, including:

[0042] Step A, obtain the previous second number of data in the ICA data as the current data;

[0043] Step B, input the current data into the ICA data acquisition model to obtain the next data of the current data;

[0044] Step C, use the next data as the last data of the second number of data, and re-obtain the second number of data as the current data;

[0045] Repeat Step B and Step C until the dQ / dV value corresponding to the upper limit voltage of the first voltage range is obtained.

[0046] Solution 5. The method for estimating the SOH of a battery according to Solution 3, characterized in that obtain the historical dQ / dV data sequence corresponding to the third voltage range obtained according to the actual sampling frequency during the charging process of the battery;

[0047] Based on the historical dQ / dV data sequence, according to the first voltage range and the voltage sampling step size, obtain a sample dQ / dV data sequence through interpolation operation;

[0048] Train the ICA data acquisition model and the battery SOH estimation model respectively based on the sample dQ / dV data sequence;

[0049] Wherein, the lower limit voltage of the third voltage range is less than or equal to the lower limit voltage of the first voltage range, and the upper limit voltage of the third voltage range is greater than or equal to the upper limit voltage of the first voltage range.

[0050] Solution 6. The battery SOH estimation method according to Solution 1, wherein obtaining the SOH estimation value of the battery includes: obtaining the SOH estimation value of the battery through a convolutional neural network.

[0051] Solution 7. The battery SOH estimation method according to Solution 6, wherein the convolutional neural network includes a LeNet network, the LeNet network sequentially includes a plurality of convolutional blocks and fully connected layers, the number of input channels of the first convolutional block among the plurality of convolutional blocks is denoted as the first number of channels, the number of output channels of the last convolutional block among the plurality of convolutional blocks is denoted as the second number of channels, and the first number of channels is less than the second number of channels.

[0052] Solution 8. The battery SOH estimation method according to Solution 3, wherein the method further includes a model training method, and the model training method includes:

[0053] Obtain the historical ICA data corresponding to the third voltage range obtained according to the actual sampling frequency during the charging process of the battery;

[0054] Based on the historical ICA data, perform interpolation operation on the historical ICA data to obtain sample ICA data;

[0055] Train the ICA data acquisition model based on the sample ICA data;

[0056] Wherein, the lower limit voltage of the third voltage range is less than or equal to the lower limit voltage of the first voltage range, and the upper limit voltage of the third voltage range is greater than or equal to the upper limit voltage of the first voltage range.

[0057] Solution 9. The battery SOH estimation method according to Solution 4, wherein the upper limit voltage of the first voltage range is the charging cut-off voltage of the battery.

[0058] Solution 10. A control device includes at least one processor and at least one storage device. The storage device is adapted to store multiple program codes. It is characterized in that the program codes are adapted to be loaded and run by the processor to execute the battery SOH estimation method described in any one of Solutions 1 to 9.

[0059] Solution 11. A storage medium is adapted to store multiple program codes. It is characterized in that the program codes are adapted to be loaded and run by a processor to execute the battery SOH estimation method described in any one of Solutions 1 to 9. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Referring to the accompanying drawings, the disclosure of the present application will become more understandable. It is easy for those skilled in the art to understand that these drawings are only for illustrative purposes and are not intended to limit the protection scope of the present application.

[0061] Figure 1 is the main flowchart of the battery SOH estimation method of the embodiments of the present application.

[0062] Figure 2 is a specific implementation manner for supplementing ICA data in step S200 of the battery SOH estimation method of the embodiments of the present application.

[0063] Figure 3 is a comparison diagram of the ICA curve obtained by the battery SOH estimation method of the embodiments of the present application and the actually measured ICA curve. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the protection scope of the present application.

[0065] Those skilled in the art should understand that these embodiments are only used to explain the technical principle of the present application and are not intended to limit the protection scope of the present application. Those skilled in the art can make adjustments according to needs to adapt to specific application scenarios.

[0066] Next, a specific implementation manner of the battery SOH estimation method mentioned in the present application will be described with reference to the accompanying drawings.

[0067] As Figures 1-3 shown, the battery SOH estimation method of the present application includes:

[0068] S100. During the charging process, obtain the ICA data of the battery;

[0069] Further, the ICA data includes a sequence of the first number of data within a first voltage range. When the first number is less than a preset number, trigger the replenishment instruction.

[0070] S200. In response to the replenishment instruction, replenish the ICA data to obtain replenished data;

[0071] Further, the ICA data is replenished through an ICA data acquisition model.

[0072] Still further, the first data in the ICA data is the dQ / dV value corresponding to the lower limit voltage of the first voltage range, and the ICA data is the first number of dQ / dV values. When the first number is greater than or equal to the second number, replenishing the ICA data through the trained ICA data acquisition model includes:

[0073] Step A. Obtain the previous second number of data in the ICA data as the current data;

[0074] Step B. Input the current data into the ICA data acquisition model to obtain the next data of the current data;

[0075] Step C. Use the next data as the last data of the second number of data, and re-obtain the second number of data as the current data;

[0076] Repeat Step B and Step C until the dQ / dV value corresponding to the upper limit voltage of the first voltage range is obtained.

[0077] S300. Obtain the SOH estimation value of the battery according to the replenished data.

[0078] Further, obtaining the SOH estimation value of the battery includes: obtaining the SOH estimation value of the battery through a convolutional neural network.

[0079] Next, for steps S100 to S300, detailed expansions will be carried out respectively in combination with specific embodiments.

[0080] In steps S100 and S200, the ICA data of the battery is obtained. The ICA data can be dQ / dV, or other solutions that can reflect IAC data disclosed in the current prior art. Taking dQ / dV as an example below, a detailed expansion will be carried out on how to obtain the ICA data and how to replenish the ICA data mentioned in S100 and S200:

[0081] In a possible implementation, the battery can be a lithium iron phosphate battery. The upper limit voltage of the first voltage range is the charging cut-off voltage of the lithium iron phosphate battery. Taking the first voltage range as [3.300, 3.650] and the voltage sampling step as 0.005V as an example for illustration. Of course, the values can be adjusted according to actual needs. For example, 3.300V can be modified to 3.280V, etc. In the conventional solutions of the prior art, it is necessary to collect all sampling points with a step of 0.005V between 3.300V and 3.650V (in this embodiment of the preset range, there are 71 sampling points, 3.300V, 3.305V, 3.310V... 3.650V), and obtain the dQ / dV values of all sampling points. Taking the dQ / dV value as the ordinate and the voltage value as the abscissa, draw a complete ICA curve (discharge curve of the battery) to reasonably estimate the battery SOH, give an accurate judgment, and avoid the user's mileage anxiety caused by inaccurate range estimation. However, if the user directly terminates the charging without fully charging, at this time, it is impossible to completely sample 71 points, which will cause incomplete sampling, the ICA curve cannot be drawn completely, and the SOH value cannot be calculated, so that the user cannot obtain the real SOH value in this application scenario and can only use the last SOH value. In this way, if the SOH value is not corrected due to long-term non-full charging, it will cause inaccurate vehicle mileage evaluation.

[0082] However, in this application, it is not necessary to fully collect. In a possible implementation of step S100, it is set that "the ICA data includes a sequence of the first number of data within the first voltage range". Taking the first number as 16 as an example, that is, among the current 71 values, only the first 16 points need to be sampled as the values of the first dQ / dV data sequence for evaluating the SOH in this application. At this time, the V values are 3.300V, 3.305V, 3.310V... 3.375V, that is, only need to sample 0.075V of vehicle charging to obtain the 16 values we need. Even if the battery is not fully charged to 3.65V (taking 0.35V of full charge as an example), this application can accurately complete the remaining ICA curve and realize the accurate evaluation of SOH. The first element in the first dQ / dV data sequence (that is, 1 in 1-16) is the dQ / dV value corresponding to the lower limit voltage of the first voltage range (that is, the dQ / dV value corresponding to 3.300V). The first number (that is, 16) is greater than or equal to the window size of the sliding time window (the window size of the sliding time window is the second number, the first number is greater than or equal to the second number, and in this embodiment, equality is used for illustration, that is, the second number is also 16).

[0083] The following is an example for steps A to C. Taking the size of the sliding time window as 16, a sequence of 16 data of the first quantity (values from 1 to 16) generates a first model input sequence, and then it is input into a trained ICA data acquisition model (such as an LSTM model) to obtain the calculated dQ / dV value corresponding to the next voltage sampling point of the first current voltage sampling point, that is, to predict the value with an index of 17. The calculated dQ / dV value (the value of 17) is added to the end of the first dQ / dV data sequence (values from 1 to 16, and the second quantity is also 16) to generate a new first dQ / dV data sequence (values from 1 to 17) composed of the first dQ / dV data sequence and the calculated dQ / dV value. Then, the window is slid once (the window value is still the second quantity of 16), and the values with indexes from 2 to 17 are re-input into the Transformer model to predict the value with an index of 18, and so on, until the values with indexes from 55 to 70 predict the value with an index of 71. The 71 dQ / dV sampling points generated from the charging data of a battery pack on a certain day can have 55 training data of sliding time windows. These 55 training data are used to complement the initially measured 16 data, and finally, 71 data sequences in the complete first voltage range [3.300, 3.650] are obtained, thereby completing the supplementation of the ICA data to obtain supplementary data.

[0084] To reduce the computational amount and increase the accuracy, when the size of the sliding time window is still 16, but the actual sampled data exceeds 16. For example, if 20 data are actually sampled, then starting to predict from 17 is a waste of calculation, and the actual accuracy will decrease because there is a measured value for 17, which is more accurate. To adapt to the actual sampled values, for example, if 20 groups of data are actually sampled, then the last element is 20, and the last element of the first dQ / dV data sequence is also 20. The first dQ / dV data sequence is then values from 5 to 20. Then, only the value of 21 needs to be deduced from the values from 5 to 20, and so on, until the values with indexes from 55 to 70 predict the value with an index of 71 (the dQ / dV value corresponding to the upper limit voltage of 3.650V in the first voltage range). This reduces four deductions of 1 - 16, 2 - 17, 3 - 18, and 4 - 19, thereby reducing the computational amount and improving the accuracy.

[0085] This application also proposes a supplementary instruction scheme. Only when the first quantity (16) is less than the preset quantity (71), the supplementary instruction is triggered and responded to, and the ICA data is supplemented. If the first quantity is equal to the preset quantity, there is no need for supplementation because the real data of the ICA curve has been collected and there is no need for further deduction.

[0086] Due to different charging piles, different initial charging voltages of vehicles, different sampling frequencies, etc., it is usually impossible to directly obtain the 16 points of V values required by this application, which are 3.300V, 3.305V, 3.310V... 3.375V. For example, when the initial voltage of the vehicle is 3.298V and the sampling frequency is 0.003V, the sampled values obtained are 3.298V, 3.301V, 3.304V, 3.307V, etc. It can be seen that it does not sample 3.300V and 3.305V. To solve this practical problem, this application also proposes a further refinement scheme, including:

[0087] Obtain the original dQ / dV data sequence corresponding to the second voltage interval obtained according to the actual sampling frequency during the charging process of the battery;

[0088] Based on the original dQ / dV data sequence, according to the first voltage interval and the voltage sampling step, obtain the first dQ / dV data sequence through interpolation operation;

[0089] Wherein, the lower limit voltage of the second voltage interval is less than or equal to the lower limit voltage of the first voltage interval.

[0090] The original dQ / dV data sequence, that is, corresponding to a possible sampling sequence mentioned in this application, such as 3.298V, 3.301V, 3.304V, 3.307V... 3.376V, 3.379V, etc. That is, when the actual sampling frequency is 0.003V, the second voltage interval at this time is [3.298, 3.376]. At this time, the sampled data of the 16 points of 3.300V, 3.305V, 3.310V... 3.375V actually required by this application can be obtained through interpolation operation. The calculation method can be to calculate after connecting points with a straight line, or to obtain the numerical value after obtaining the fitting curve. The acquisition methods are diverse, and those skilled in the art can actually select according to needs. In order to ensure that the numerical values of these 16 points in the range of [3.300, 3.375] of this application must be complete, the value range of the second voltage interval is specified. The lower limit voltage of the second voltage interval needs to be less than or equal to the lower limit voltage of the first voltage interval, that is, 3.298V < 3.300V, to ensure the integrity of the starting point. The end point is different due to different selected first quantities (16 in the embodiment), so there is no need to limit it, and it only needs to be obtained according to the actual situation later.

[0091] Above, in combination with a specific scheme, the acquisition method of the ICA curve has been expanded. Among them, the ICA data acquisition model is applied. Next, the training method of the ICA data acquisition model will be expanded:

[0092] The ICA data acquisition model mentioned in this application can derive a complete dQ / dV data sequence (e.g., 71) based on a preset first dQ / dV data sequence (e.g., 16), and then obtain a completed ICA model (the first 16 are obtained through actual acquisition, and the last 55 are obtained through derivation). Finally, the ICA curve corresponding to the first voltage range is obtained. There are various ICA data acquisition models that can be used for the derivation, such as the LSTM model, the recurrent network RNN, GRU, etc. Among them, the prediction is carried out through a trained ICA data acquisition model. This application constructs the ICA data acquisition model based on a recurrent neural network, preferably implemented by the LSTM model, and can also be implemented using Transformer, SRU, RQNN, etc. As Figure 3 shown, the actual data and the predicted data of the same battery pack are compared. The coincidence degree of the dotted line and the solid line is very high, making the average error of SOH within 0.5%, fully meeting the accuracy requirements for SOH estimation.

[0093] After the supplementary data is obtained in step S200, step S300 will be elaborated in detail below:

[0094] S300. Obtain the SOH estimated value of the battery according to the supplementary data.

[0095] Furthermore, obtaining the SOH estimated value of the battery includes: obtaining the SOH estimated value of the battery through a convolutional neural network.

[0096] Furthermore, the convolutional neural network includes a LeNet network. The LeNet network sequentially includes a plurality of convolutional blocks and fully connected layers. The number of input channels of the first convolutional block among the plurality of convolutional blocks is denoted as the first number of channels, and the number of output channels of the last convolutional block among the plurality of convolutional blocks is denoted as the second number of channels. The first number of channels is less than the second number of channels.

[0097] First, the supplementary data needs to be converted into a picture format, that is, an ICA curve image is obtained based on the completed dQ / dV data sequence corresponding to the first voltage range, and then the image is input into the convolutional neural network for processing to obtain the SOH value;

[0098] Specifically, a possible implementation manner of how to convert the dQ / dV data sequence into an image is as follows:

[0099] Establish an ICA curve image template, and the ICA curve image template includes:

[0100] The image size of the unified ICA curve image template, the origin position of the unified ICA curve coordinate system origin in the ICA curve image template, the abscissa unit length of the unified ICA curve, and the ordinate unit length of the unified ICA curve;

[0101] Based on the dQ / dV data sequence corresponding to the first voltage interval, draw an ICA curve in the ICA curve image template to obtain the first ICA curve image;

[0102] Hide the abscissa and ordinate in the first ICA curve image to obtain the ICA curve image.

[0103] First, design a set of unified image templates, then draw in the image template to obtain the first ICA curve image. After the drawing is completed, hide the abscissa and ordinate in the image. Since the SOH value is obtained through image processing in the later stage of this application, in order to avoid more interference and improve accuracy, a scheme for excluding redundant interference of the ICA curve is designed. Since the coordinate system and other structures are all unified, and finally the coordinate system is hidden, it can maximize the avoidance of image calculation and analysis errors caused by redundant visible features.

[0104] After obtaining the ICA curve image, continue to obtain the SOH value of the battery:

[0105] Obtain the SOH estimated value of the battery through a convolutional neural network.

[0106] The battery SOH estimation model can be constructed based on a convolutional neural network. In a possible implementation, the convolutional neural network is a LeNet network. The LeNet network sequentially includes multiple convolutional blocks and a fully connected layer. The number of input channels of the first convolutional block among the multiple convolutional blocks is denoted as the first number of channels, and the number of output channels of the last convolutional block among the multiple convolutional blocks is denoted as the second number of channels. The first number of channels is less than the second number of channels. Further, the LeNet network sequentially includes a first convolutional block, a second convolutional block, a third convolutional block, a fourth convolutional block, and a fully connected layer. The output of the fully connected layer is the SOH estimated value of the battery. Among them, the first convolutional block sequentially includes a first convolutional kernel, a first BN layer, a first activation function layer, and a first pooling layer; the second convolutional block sequentially includes a second convolutional kernel, a second BN layer, a second activation function layer, and a second pooling layer; the third convolutional block sequentially includes a third convolutional kernel, a third BN layer, a third activation function layer, and a third pooling layer; the fourth convolutional block sequentially includes a fourth convolutional kernel, a fourth BN layer, a fourth activation function layer, and a fourth pooling layer; the input of the LeNet network is the original data of the ICA curve image in the first dimension with the first number of channels, and the output of the fourth convolutional block is the feature data of the ICA curve image in the second dimension with the second number of channels. Among them, the first number of channels is less than the second number of channels, and the first dimension is greater than the second dimension.

[0107] The combined effect of the convolutional layer and the pooling layer increases the number of channels (number of features) and reduces the image size. The convolutional neural network can input the curve image of ICA and output the estimated value of SOH after calculation. Since calculating the estimated value of SOH based on the known complete ICA curve image is a common means that those skilled in the art can easily obtain, the scheme of training LeNet based on this does not need to be elaborated in detail here. The core invention here is to apply LeNet to this step for the complete implementation of the SOH estimation method, rather than the parameter modification of LeNet itself.

[0108] So far, the overall scheme uses two models, namely the ICA data acquisition model constructed by the recurrent neural network LSTM and the battery SOH estimation model for analyzing images constructed by the convolutional neural network LeNet. Both of these models need to be trained with a data sample set. The data sample set acquisition method is as follows:

[0109] Obtain the historical dQ / dV data sequence corresponding to the third voltage interval obtained according to the actual sampling frequency during the battery charging process;

[0110] Based on the historical dQ / dV data sequence, obtain the sample dQ / dV data sequence through interpolation operation according to the first voltage interval and the voltage sampling step size;

[0111] Train the ICA data acquisition model and the battery SOH estimation model respectively based on the sample dQ / dV data sequence;

[0112] Among them, the lower limit voltage of the third voltage interval is less than or equal to the lower limit voltage of the first voltage interval, and the upper limit voltage of the third voltage interval is greater than or equal to the upper limit voltage of the first voltage interval.

[0113] With such settings, the third voltage interval can completely enclose the first voltage interval, so that the sampling data of the first voltage interval is complete, ensuring that the predictions of the ICA data acquisition model and the SOH estimation model in the first voltage interval will not be distorted at the edges, and making the model trained by the samples more accurate.

[0114] Of course, obtaining historical data is not limited to the historical dQ / dV data sequence, and it can also be directly obtaining historical ICA data. Correspondingly, the above data sample set acquisition method can also be as follows:

[0115] Obtain the historical ICA data corresponding to the third voltage interval obtained according to the actual sampling frequency during the battery charging process;

[0116] Based on the historical ICA data, perform interpolation operation on the historical ICA data to obtain sample ICA data;

[0117] Train the ICA data acquisition model based on the sample ICA data;

[0118] Wherein, the lower voltage of the third voltage range is less than or equal to the lower voltage of the first voltage range, and the upper voltage of the third voltage range is greater than or equal to the upper voltage of the first voltage range.

[0119] Finally, in the present application, by only measuring the values of some preset charging times (such as 16, 17, 18, etc.), a complete ICA curve can be obtained, and finally an accurate SOH value can be obtained without waiting for the battery to be fully charged, expanding the application scenario of SOH prediction and enabling users to further distance themselves from range anxiety.

[0120] Furthermore, the present application also provides a battery SOH estimation system.

[0121] The battery SOH estimation system in the embodiments of the present application mainly includes:

[0122] ICA data acquisition module: used to acquire the ICA data of the battery during the charging process;

[0123] Supplementary data module: used to supplement the ICA data in response to a supplementary instruction to obtain supplementary data;

[0124] SOH estimation module: used to obtain the SOH estimated value of the battery according to the supplementary data.

[0125] In some embodiments, one or more of the above three modules can be combined into one module.

[0126] It should be noted that the ordinal numbers such as "first" and "second" in the description and claims of the present application and the above-mentioned drawings are only used to distinguish similar objects, rather than to describe or represent a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here.

[0127] Furthermore, the present application also provides a control device. In an embodiment of a control device according to the present application, the control device includes a processor and a storage device. The storage device can be configured to store a program for executing the battery SOH estimation method in the above method embodiments, and the processor can be configured to execute the program in the storage device. The program includes, but is not limited to, a program for executing the battery SOH estimation method in the above method embodiments. For the sake of convenience of description, only the parts related to the embodiments of the present application are shown. For the specific technical details not disclosed, please refer to the method part of the embodiments of the present application. The control device can be a control device formed by various electronic devices.

[0128] In the embodiments of the present application, the control device may be a control device formed by various electronic devices. In some possible implementation manners, the control device may include a plurality of storage devices and a plurality of processors. The program for executing the battery SOH estimation method in the above method embodiments may be divided into multiple sub-programs, and each sub-program may be loaded and run by a processor respectively to execute different steps of the battery SOH estimation method in the above method embodiments. Specifically, each sub-program may be stored in a different storage device respectively, and each processor may be configured to execute the program in one or more storage devices to jointly implement the battery SOH estimation method in the above method embodiments, that is, each processor executes different steps of the battery SOH estimation method in the above method embodiments to jointly implement the battery SOH estimation method in the above method embodiments.

[0129] The above-mentioned plurality of processors may be processors deployed on the same device. For example, the above-mentioned control device may be a high-performance device composed of a plurality of processors, and the above-mentioned plurality of processors may be processors configured on the high-performance device. In addition, the above-mentioned plurality of processors may also be processors deployed on different devices. For example, the above-mentioned control device may be a server cluster, and the above-mentioned plurality of processors may be processors on different servers in the server cluster.

[0130] Furthermore, the present application also provides a computer-readable storage medium. In an embodiment of a computer-readable storage medium according to the present application, the computer-readable storage medium may be configured to store a program for executing the battery SOH estimation method in the above method embodiments, and this program may be loaded and run by a processor to implement the above battery SOH estimation method. For the sake of convenience of description, only the parts related to the embodiments of the present application are shown. For the specific technical details not disclosed, please refer to the method part of the embodiments of the present application. The computer-readable storage medium may be a storage device formed by various electronic devices. Optionally, the computer-readable storage medium in the embodiments of the present application is a non-transitory computer-readable storage medium.

[0131] Furthermore, it should be understood that since the setting of each module is only for explaining the functional units of the device of the present application, the physical devices corresponding to these modules may be the processor itself, or a part of the software in the processor, a part of the hardware, or a part of the combination of software and hardware. Therefore, the number of each module in the figure is only illustrative.

[0132] Those skilled in the art can understand that the various modules in the device can be adaptively split or combined. Such splitting or combination of specific modules will not cause the technical solution to deviate from the principle of the present application. Therefore, the technical solutions after splitting or combination will all fall within the protection scope of the present application.

[0133] In each embodiment of this application, any relevant personal information of users involved may be processed in strict accordance with the requirements of laws and regulations, following the principles of legality, legitimacy, and necessity, for reasonable purposes based on business scenarios, and is personal information actively provided by users during the use of products / services, generated due to the use of products / services, or obtained with user authorization.

[0134] The personal information of users processed in this application may vary depending on the specific product / service scenario. It is subject to the specific scenario of the user's use of the product / service and may involve the user's account information, device information, driving information, vehicle information, or other relevant information. The applicant will treat the user's personal information and its processing with a high degree of diligence.

[0135] This application attaches great importance to the security of users' personal information and has taken security protection measures that meet industry standards and are reasonable and feasible to protect users' information, preventing personal information from being accessed without authorization, publicly disclosed, used, modified, damaged, or lost.

[0136] It should be noted that the ordinal numbers such as "first" and "second" in the description, claims, and the above-mentioned drawings of this application are only used to distinguish similar objects, rather than to describe or represent a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here.

[0137] So far, the technical solutions of this application have been described in combination with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of this application is obviously not limited to these specific embodiments. Without departing from the principle of this application, those skilled in the art can make equivalent changes or substitutions to relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of this application.

Claims

1. A method for estimating the state of health (SOH) of a battery, characterized in that, The method includes: During the charging process, obtaining ICA data of the battery; In response to a supplementary instruction, supplementing the ICA data to obtain supplementary data; Based on the supplementary data, obtaining an SOH estimation value of the battery.

2. The battery SOH estimation method according to claim 1, wherein Supplementing the ICA data includes: Supplementing the ICA data through an ICA data acquisition model.

3. The battery SOH estimation method according to claim 1 or 2, characterized in that The ICA data includes a sequence of a first number of data within a first voltage range. When the first number is less than a preset number, the supplementary instruction is triggered.

4. The battery SOH estimation method according to claim 3, characterized in that, The first data in the ICA data is the dQ / dV value corresponding to the lower limit voltage of the first voltage range. The ICA data is a first number of dQ / dV values. When the first number is greater than or equal to a second number, supplementing the ICA data through an ICA data acquisition model includes: Step A, obtaining the previous second number of data in the ICA data as current data; Step B, inputting the current data into the ICA data acquisition model to obtain the next data of the current data; Step C, using the next data as the last data of the second number of data, and re-obtaining the second number of data as current data; Repeatedly execute Step B and Step C until the dQ / dV value corresponding to the upper limit voltage of the first voltage range is obtained.

5. The battery SOH estimation method according to claim 3, wherein Obtaining a historical dQ / dV data sequence corresponding to a third voltage range obtained according to the actual sampling frequency during the battery charging process; Based on the historical dQ / dV data sequence, obtaining a sample dQ / dV data sequence through interpolation operation according to the first voltage range and the voltage sampling step size; Based on the sample dQ / dV data sequence, training the ICA data acquisition model and the battery SOH estimation model respectively; Wherein, the lower limit voltage of the third voltage range is less than or equal to the lower limit voltage of the first voltage range, and the upper limit voltage of the third voltage range is greater than or equal to the upper limit voltage of the first voltage range.

6. The battery SOH estimation method according to claim 1, characterized in that, Obtaining the SOH estimation value of the battery includes: obtaining the SOH estimation value of the battery through a convolutional neural network.

7. The battery SOH estimation method according to claim 6, wherein The convolutional neural network includes a LeNet network. The LeNet network sequentially includes a plurality of convolutional blocks and fully connected layers. The number of input channels of the first convolutional block in the plurality of convolutional blocks is denoted as the first number of channels, and the number of output channels of the last convolutional block in the plurality of convolutional blocks is denoted as the second number of channels. The first number of channels is less than the second number of channels.

8. The battery SOH estimation method according to claim 3, wherein The method further includes a model training method, and the model training method includes: Obtaining historical ICA data corresponding to a third voltage range obtained according to the actual sampling frequency during the battery charging process; Based on the historical ICA data, performing interpolation operation on the historical ICA data to obtain sample ICA data; Based on the sample ICA data, training the ICA data acquisition model; Wherein, the lower limit voltage of the third voltage range is less than or equal to the lower limit voltage of the first voltage range, and the upper limit voltage of the third voltage range is greater than or equal to the upper limit voltage of the first voltage range.

9. The battery SOH estimation method according to claim 4, characterized in that, The upper limit voltage of the first voltage range is the charging cut-off voltage of the battery.

10. A control device, comprising at least one processor and at least one storage device, the storage device being adapted to store a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by the processor to execute the battery SOH estimation method according to any one of claims 1 to 9.

Images