Battery state of health calculation method, device, storage medium, and vehicle

Abstract

The application provides a battery state of health calculation method and device, a storage medium and a vehicle, and belongs to the technical field of data processing. The method comprises the following steps: obtaining historical discharge data of a battery, and preprocessing the obtained historical discharge data; inputting the preprocessed historical discharge data into a state calculation model to obtain a current battery state of health; the state calculation model comprises a first calculation network and a second calculation network which are obtained by joint training based on a historical sample data set; the first calculation network identifies a degradation degree based on the historical sample data set; and the second calculation network combines the degradation degree identification result of the first calculation network and the historical sample data set to output a health state identification result of each battery sample. The application aims to obtain a more accurate battery state of health calculation result.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to a method, apparatus, storage medium, and vehicle for calculating battery health status. Background Technology

[0002] In battery management systems, State of Health (SOH) is a very important battery parameter. As a quantitative indicator for evaluating the degree of battery aging, SOH affects vehicle strategies and the extent of battery usage. An accurate SOH can effectively improve battery performance and battery life, thereby enhancing the driver's experience.

[0003] Battery health status (SOH) is affected by multiple factors, including the depth of discharge. However, there is no non-linear relationship between the depth of discharge, the number of discharges, and the SOH value. If the SOH value is calculated directly based on the depth of discharge, the calculated SOH value will have a large error and poor accuracy. Summary of the Invention

[0004] This application provides a battery health state calculation method, apparatus, storage medium, and vehicle, aiming to obtain more accurate battery health state calculation results.

[0005] In a first aspect, embodiments of this application provide a method for calculating battery health status, the method comprising:

[0006] Acquire historical discharge data of the battery under test, including battery temperature, discharge factor, depth of discharge, and number of depths of discharge during different usage stages of the battery under test.

[0007] The historical discharge data of the battery under test is preprocessed.

[0008] The preprocessed historical discharge data is input into the state calculation model to obtain the current battery health status of the battery under test.

[0009] The state calculation model includes a first calculation network and a second calculation network jointly trained based on a historical sample dataset. The historical sample dataset includes battery temperature samples, discharge rate samples, discharge depth samples, and depth counts of multiple battery samples at different usage stages. The first calculation network identifies the degradation degree of the battery samples, and the second calculation network combines the degradation degree identification result output by the first calculation network with the historical sample dataset to output the health status identification result of each of the multiple battery samples.

[0010] Optionally, the depth of discharge is represented by a gradient, which is determined based on the remaining capacity of the battery under test at the last moment of each usage phase.

[0011] Optionally, the gradient is determined based on the remaining capacity of the battery under test at the last moment of each usage phase, including:

[0012] Obtain the remaining capacity of the battery under test at the last moment of each usage phase;

[0013] Among multiple preset remaining capacity ranges, the remaining capacity range to which the remaining capacity belongs is determined, and the gradient corresponding to the remaining capacity range is used as the gradient of the depth of discharge of the battery under test in this usage stage.

[0014] The multiple preset remaining capacity intervals are intervals obtained by dividing the remaining capacity values ​​according to preset intervals, and each remaining capacity interval corresponds to a different gradient.

[0015] Optionally, the state calculation model is trained according to the following steps:

[0016] The historical sample dataset is constructed by acquiring historical sample data of multiple battery samples at different usage stages and the current real battery health status of each of the multiple battery samples. The historical sample data includes: battery temperature samples, discharge rate samples, discharge depth samples and depth counts.

[0017] The historical sample dataset is preprocessed;

[0018] Construct the first computing network and the second computing network respectively;

[0019] The first computing network, taking the preprocessed historical sample dataset as input, identifies the degradation degree of each battery sample among the multiple battery samples;

[0020] The second computing network takes the preprocessed historical sample dataset and the identification result of the degradation degree of each battery sample by the first computing network as input, and outputs the identification result of the battery health status of each battery sample among the multiple battery samples;

[0021] Based on the battery health status identification results of each battery sample output by the second computing network, the parameters of the first computing network and the second computing network are updated.

[0022] Optionally, both the first computing network and the second computing network include an input layer, a hidden layer, and an output layer, wherein,

[0023] The transfer function from the input layer to the hidden layer is:

[0024]

[0025] The transfer function from the hidden layer to the output layer is:

[0026] f2(x)=x.

[0027] Optionally, the formula for calculating the j-th neuron in the hidden layer is:

[0028]

[0029] In the formula, w ij x represents the weights from the i-th neuron in the input layer to the j-th neuron in the hidden layer. i Let b be the value of the i-th neuron in the input layer. j is the threshold of the j-th neuron in the hidden layer, and I is the total number of neurons in the input layer;

[0030] From the hidden layer to the output layer, the formula for calculating the k-th neuron in the output layer is:

[0031]

[0032] In the formula, M j w is the value of the j-th neuron in the hidden layer. jk Let a be the weights from the j-th neuron in the hidden layer to the k-th neuron in the output layer. k J is the threshold of the k-th neuron in the output layer, and J is the total number of neurons in the hidden layer.

[0033] Optionally, based on the battery health status identification results of each battery sample output by the second computing network, the parameters of the first computing network and the second computing network are updated, including:

[0034] Based on the identification result of the battery health status of each battery sample output by the second computing network, the error between the identification result and the current true battery health status of each battery sample is determined.

[0035] When the error is greater than or equal to the calibration value, the parameters of the first computing network and the second computing network are iteratively updated.

[0036] When the error is less than the calibration value, the weights and thresholds from the input layer to the hidden layer of the first computing network and the weights and thresholds from the hidden layer to the output layer of the second computing network are used as the model parameters of the trained state computing model.

[0037] Secondly, embodiments of this application provide a battery health status calculation device, the device comprising:

[0038] The data acquisition module is used to acquire historical discharge data of the battery under test. The historical discharge data includes the battery temperature, discharge factor, discharge depth and number of discharges of the battery under test in different usage stages.

[0039] A preprocessing module is used to preprocess the historical discharge data of the acquired battery under test;

[0040] The state calculation module is used to input the preprocessed historical discharge data into the state calculation model to obtain the current battery health state of the battery under test.

[0041] The state calculation model includes a first calculation network and a second calculation network jointly trained based on a historical sample dataset. The historical sample dataset includes battery temperature samples, discharge rate samples, discharge depth samples, and depth counts of multiple battery samples at different usage stages. The first calculation network identifies the degradation degree of the battery samples, and the second calculation network combines the degradation degree identification result output by the first calculation network with the historical sample dataset to output the health status identification result of each of the multiple battery samples.

[0042] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the battery health state calculation method as described in the first aspect of the embodiments.

[0043] Fourthly, this application provides a vehicle equipped with a battery health status calculation device as described in the second aspect of the embodiment, the battery health status calculation device executing the battery health status calculation method as described in the first aspect of the embodiment.

[0044] Beneficial effects:

[0045] The state calculation model in this method includes a two-layer neural network, namely a first calculation network and a second calculation network. The first calculation network is trained based on multiple battery samples at different usage stages, including battery temperature samples, discharge rate samples, discharge depth samples, and depth counts. It can identify the relationship between these factors and the degradation degree, and thus output a more accurate degradation degree prediction. When the second calculation network predicts the current battery health state, it calculates based on the degradation degree identification results output by the first calculation network, which can obtain a more accurate battery health state calculation result. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of the steps of a battery health status calculation method proposed in an embodiment of this application;

[0048] Figure 2 This is a flowchart of the steps of a training state calculation model proposed in an embodiment of this application;

[0049] Figure 3 This is a functional block diagram of a battery health status calculation device proposed in an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0051] Battery State of Health (SOH) is crucial for vehicle-related strategies and battery safety. For example, if the current SOH value of the battery is insufficient to support the execution of a function when the vehicle performs a function, the function cannot be executed successfully. If the function is forced to be executed, it may lead to a rapid reduction in battery life. Therefore, accurate SOH value is very important to improve battery performance and battery life.

[0052] Factors affecting the SOH value of a battery include at least: battery temperature, discharge rate, depth of discharge and number of discharges, and SOH decay rate. However, there is no non-linear relationship between the depth of discharge and the number of discharges and the SOH value. If the SOH value is calculated directly based on the depth of discharge and the number of discharges, the result will have a large error.

[0053] To obtain more accurate battery health state calculation results, this application proposes a battery health state calculation method.

[0054] Reference Figure 1 The diagram illustrates a flowchart of a battery health status calculation method according to an embodiment of the present invention. The method may specifically include the following steps:

[0055] S101: Obtain historical discharge data of the battery under test, including battery temperature, discharge factor, discharge depth and number of discharges of the battery under test in different usage stages.

[0056] When calculating the current state of health (SOH) of the battery under test, historical discharge data of the battery under test in different stages of use to date is obtained, including battery temperature, discharge factor, depth of discharge, and number of depth discharges.

[0057] In practice, different usage stages can be determined according to the actual situation. For example, each use from the start of using the battery under test to the turn off of the battery under test can be set as a usage stage, or the usage status of the battery under test within a preset time period can be used as a usage stage, such as the usage status of the vehicle's battery under test within a day.

[0058] S102: Preprocess the historical discharge data of the battery under test obtained.

[0059] Preprocessing specifically refers to normalizing the historical discharge data using a linear function to eliminate dimensions, so that each historical discharge data falls within the range of [0, 1].

[0060] S103: Input the preprocessed historical discharge data into the state calculation model to obtain the current battery health status of the battery under test.

[0061] The state calculation model calculates the current state of battery health, i.e., the current SOH value, based on the historical discharge data of the battery under test in different stages of use, given the initial SOH value of the battery.

[0062] The state calculation model includes a two-layer neural network, specifically a first calculation network and a second calculation network jointly trained based on the historical sample dataset. The historical sample dataset includes battery temperature samples, discharge rate samples, discharge depth samples, and depth counts of multiple battery samples at different usage stages.

[0063] The first computing network identifies the decay rate based on the historical sample dataset. Since the first computing network is trained on battery temperature samples, discharge rate samples, discharge depth samples and depth counts of multiple battery samples at different usage stages, it can identify the relationship between these factors and the decay rate of SOH.

[0064] In particular, since there is no nonlinear relationship between the depth of discharge and the number of discharges and the SOH value, but there is a nonlinear relationship between the depth of discharge and the number of discharges and the attenuation of SOH, for example, when the depth of discharge is 80%, the attenuation is 0.02%, and when the depth of discharge is 50%, the attenuation is 0.05%, this method focuses on considering the relationship between various factors and the attenuation through the first computational network, and can obtain more accurate attenuation identification results.

[0065] The second computing network combines the attenuation identification results output by the first computing network with the historical sample dataset to obtain more accurate health status identification results for each of the multiple battery samples.

[0066] Specifically, since the SOH value of the battery gradually decreases, the relationship between the SOH value and the SOH decay rate is as follows: the SOH decay rate of the battery at time k is the difference between the SOH value at time k-1 and the SOH value at time k, that is:

[0067] SOH k，衰减度 =SOH k-1 -SOH k

[0068] Furthermore, by using historical discharge data of the battery at different stages of use, the first calculation network first obtains the accurate state of health (SOH) attenuation. Then, the second calculation network, based on the historical discharge data and the SOH attenuation, can obtain a more accurate calculation result of the current battery health status. The accurate battery health status calculation result can effectively improve battery performance and battery lifespan.

[0069] In practice, since the depth of discharge cannot be obtained directly, the depth of discharge data can be determined based on the battery's remaining capacity (State of Charge, SOC). Specifically, the depth of discharge can be represented by a gradient, which is determined based on the remaining capacity of the battery under test at the last moment of each usage stage.

[0070] In one implementation, by determining the gradient of the depth of discharge for each usage stage, the remaining capacity of the battery under test at the last moment of each usage stage can be obtained; among a plurality of preset remaining capacity intervals, the remaining capacity interval to which the remaining capacity belongs is determined, and the gradient corresponding to the remaining capacity interval is used as the gradient of the depth of discharge of the battery under test in that usage stage; wherein, the plurality of preset remaining capacity intervals are intervals obtained by dividing the value of the remaining capacity according to preset intervals, and each remaining capacity interval corresponds to a different gradient.

[0071] For example, the gradient of the discharge depth can be set as follows:

[0072] When the remaining capacity of the battery under test is 100% to 90%, the discharge depth is determined to be the first gradient, denoted as discharge depth A;

[0073] When the remaining capacity of the battery under test is 90% to 80%, the discharge depth is determined to be the second gradient, denoted as discharge depth B; and so on.

[0074] Characterizing the depth of discharge with the remaining charge (SOC) as different gradients, and using the gradient form for the depth of discharge samples in the state calculation model, can reduce the complexity of iteration during model training and make the model's identification results more likely to converge.

[0075] Reference Figure 2 The flowchart illustrating the steps of training the state calculation model according to an embodiment of this application is shown. In one feasible implementation, the state calculation model is trained according to the following steps:

[0076] A1: Obtain historical sample data of multiple battery samples at different usage stages, as well as the current real battery health status of each of the multiple battery samples, and construct the historical sample dataset.

[0077] When training the state calculation model, historical sample data of multiple battery samples in different usage stages are obtained, as well as the current real battery health status of each of the multiple battery samples, to construct a historical sample dataset. The state calculation model will process the historical sample data of each battery sample in the multiple battery samples separately.

[0078] The historical sample data includes: battery temperature samples, discharge rate samples, discharge depth samples, and depth counts.

[0079] A2: Preprocess the historical sample dataset.

[0080] Preprocessing refers to normalizing different types of data in a historical sample dataset using a linear function, so that different types of data all fall within the range of [0, 1], thereby eliminating dimensions.

[0081] In one implementation, the linear functions corresponding to different types of data in the historical sample dataset are different, specifically:

[0082] The linear function corresponding to the battery temperature sample is:

[0083]

[0084] The linear function corresponding to the discharge rate sample is:

[0085]

[0086] The linear function corresponding to the number of discharge depth samples is:

[0087]

[0088] The gradient of the depth of discharge sample and the linear function of the current actual battery health state are:

[0089]

[0090] A3: Construct the first computing network and the second computing network respectively.

[0091] In this embodiment, both the first computing network and the second computing network are backpropagation (BP) neural networks, and both include an input layer, a hidden layer, and an output layer.

[0092] The transfer function from the input layer to the hidden layer is:

[0093]

[0094] In this formula, x represents the neuron value output by the input layer.

[0095] The transfer function from the hidden layer to the output layer is:

[0096] f2(x)=x

[0097] In this formula, x is the neuron value output by the hidden layer.

[0098] Since the input of the state calculation model includes the battery temperature sample, discharge rate sample, discharge depth sample and depth count for each battery sample, as well as the current real battery health state of each battery sample, in this embodiment, the input layer of both the first calculation network and the second calculation network is set with 5 neurons.

[0099] The first computational network is used to predict the decay rate of SOH, and the second computational network is used to predict the SOH value. Therefore, the output layer of both the first and second computational networks has one neuron.

[0100] The number of neurons in the hidden layer is calculated using the following formula:

[0101]

[0102] In the formula, m is the number of neurons in the input layer, n is the number of neurons in the output layer, and a is a preset value, usually 5. In this formula, the integer part of the square root operation can be taken.

[0103] In this embodiment, the number of neurons in the input layer is m = 5, the number of neurons in the output layer is n = 1, and a is 5. Therefore, the number of neurons M in the hidden layer is:

[0104]

[0105] The formula for calculating the j-th neuron in the hidden layer, from the input layer to the hidden layer, is:

[0106]

[0107] In the formula, w ij x represents the weights from the i-th neuron in the input layer to the j-th neuron in the hidden layer. i Let b be the value of the i-th neuron in the input layer. j I is the threshold of the j-th neuron in the hidden layer, and I is the total number of neurons in the input layer. In this embodiment, I = 5.

[0108] From the hidden layer to the output layer, the formula for calculating the k-th neuron in the output layer is:

[0109]

[0110] In the formula, M j w is the value of the j-th neuron in the hidden layer. jk Let a be the weights from the j-th neuron in the hidden layer to the k-th neuron in the output layer. k J is the threshold of the k-th neuron in the output layer, and J is the total number of neurons in the hidden layer. In this embodiment, J = 7 and k = 1.

[0111] In this embodiment, one hidden layer is set. If the training result does not converge, the number of hidden layers or the number of neurons in the hidden layer can be increased.

[0112] A4: The first computing network, using the preprocessed historical sample dataset as input, identifies the degradation degree of each battery sample among the multiple battery samples.

[0113] Using the preprocessed historical sample dataset as input, for a battery sample there are historical sample data from different usage stages. The first computing network outputs the SOH decay rate based on the battery temperature sample, discharge rate sample, discharge depth sample and depth count of each battery sample in different usage stages.

[0114] A5: The second computing network takes the preprocessed historical sample dataset and the identification result of the degradation degree of each battery sample by the first computing network as input, and outputs the identification result of the battery health status of each battery sample among the multiple battery samples.

[0115] In one implementation, when the first and second computing networks are trained on historical sample data of any battery sample, they sequentially input historical sample data of the battery sample at different usage stages for iterative processing. For example, the battery sample has historical sample data corresponding to the first discharge stage, the second discharge stage, and the third discharge stage in chronological order. When these historical sample data are input into the first and second computing networks, the historical sample data of the first discharge stage is input first for processing. Then, based on the processing results of the first discharge stage, the historical sample data of the second discharge stage is input, and finally the historical sample data of the third discharge stage is input. In this way, each discharge process can be considered as a unit, taking into account the influence of discharge depth, battery temperature, and discharge rate on the decay degree or SOH during each discharge process. It can also learn the cumulative effect of discharge depth at different discharge stages, so that the trained state calculation model can output more accurate battery health state calculation results.

[0116] Since battery temperature has a relatively small impact on battery degradation or state of equilibrium (SOH), in practical applications, the average value of multiple battery temperature samples from different usage stages can be used as input data to reduce data processing costs.

[0117] Since there is no nonlinear relationship between the depth of discharge (DOH) and the state of equilibrium (SOH), directly outputting the SOH value using historical sample data as input to a neural network would lead to difficulty in convergence and significant errors in the predicted SOH value. Therefore, this method utilizes a first computational network to identify the relationship between the DOH depth and its number of discharges and the SOH decay rate. The SOH decay rate identification result output by the first computational network, along with the historical sample dataset, is used as input to a second computational model. This allows for the prediction of the SOH value by considering factors such as battery temperature, discharge rate, discharge depth and its number of discharges, and the SOH decay rate.

[0118] A6: Based on the battery health status identification results of each battery sample output by the second computing network, update the parameters of the first computing network and the second computing network.

[0119] This step specifically includes the following sub-steps:

[0120] A61: Based on the identification result of the battery health status of each battery sample output by the second computing network, determine the error between the identification result and the current true battery health status of each battery sample.

[0121] In the actual training process, the historical sample dataset includes historical sample data of multiple battery samples under different usage stages. When the historical sample dataset is input into the first computing network and the second computing network, the first computing network and the second computing network process the data sample corresponding to one battery sample at a time. At this time, the second computing network outputs the SOH prediction value of the battery sample and calculates the error between the predicted value of the SOH of the battery sample and the current true battery health state (i.e. the current true SOH value) of the battery sample.

[0122] A62: When the error is greater than or equal to the calibration value, continue to iteratively update the parameters of the first computing network and the second computing network.

[0123] Specifically, if the calculated error is greater than or equal to the calibration value, the parameters of the first and second computing networks continue to be iteratively updated. In the next iteration update, the first and second computing networks can process historical sample data of another battery sample in different usage stages.

[0124] A63: When the error is less than the calibration value, the weights and thresholds from the input layer to the hidden layer of the first computing network and the second computing network, as well as the weights and thresholds from the hidden layer to the output layer, are used as the model parameters of the trained state computing model.

[0125] Specifically, if the calculated error is less than the calibration value, the first and second computational networks are trained. The trained first and second computational networks are used as the state computation model, and the parameters of the state computation model are the weights and thresholds from the input layer to the hidden layer of the first and second computational networks, and the weights and thresholds from the hidden layer to the output layer of the first and second computational networks, respectively.

[0126] This application has at least the following effects:

[0127] 1. By using the first calculation network, the effects of battery temperature, discharge rate, discharge depth and number of discharges on the decay rate of SOH are considered to obtain a more accurate decay rate of SOH;

[0128] 2. By using the identification result of SOH decay output by the first calculation network as the input of the second calculation network, and combining it with battery temperature, discharge rate, discharge depth and number of depths, a more accurate SOH value can be obtained.

[0129] 3. Since the parameters of depth of discharge are limited, the gradient of depth of discharge is determined based on the remaining capacity (SOC) of the battery, thereby avoiding the problem of inaccurate SOH calculation due to insufficient parameters.

[0130] Reference Figure 3The diagram illustrates a functional block diagram of a battery health status calculation device according to an embodiment of the present invention. The device includes:

[0131] The data acquisition module 100 is used to acquire historical discharge data of the battery under test. The historical discharge data includes the battery temperature, discharge factor, discharge depth and number of discharge depths of the battery under test in different usage stages.

[0132] The preprocessing module 200 is used to preprocess the historical discharge data of the acquired battery under test;

[0133] The state calculation module 300 is used to input the preprocessed historical discharge data into the state calculation model to obtain the current battery health state of the battery under test. The state calculation model includes a first calculation network and a second calculation network jointly trained based on a historical sample dataset. The historical sample dataset includes battery temperature samples, discharge rate samples, discharge depth samples, and depth counts for multiple battery samples at different usage stages. The first calculation network identifies the degradation degree of the battery samples, and the second calculation network combines the degradation degree identification result output by the first calculation network with the historical sample dataset to output the identification result of the health state of each of the multiple battery samples.

[0134] Optionally, the apparatus further includes a model training module, the model training module comprising:

[0135] The first acquisition unit is used to acquire historical sample data of multiple battery samples at different usage stages, as well as the current actual battery health status of each of the multiple battery samples, to construct the historical sample dataset. The historical sample data includes: battery temperature samples, discharge rate samples, discharge depth samples, and the number of discharge depths.

[0136] The first processing unit is used to preprocess the historical sample dataset;

[0137] Construction units are used to construct the first computing network and the second computing network, respectively.

[0138] The attenuation calculation unit is used to identify the attenuation of each battery sample among the plurality of battery samples by taking the preprocessed historical sample dataset as input through the first calculation network.

[0139] A battery health status prediction unit is used to take the preprocessed historical sample dataset and the degradation degree identification result of the first computing network for each battery sample as input, and output the battery health status identification result of each battery sample among the multiple battery samples.

[0140] The parameter update unit is used to update the parameters of the first computing network and the second computing network based on the identification results of the battery health status of each battery sample output by the second computing network.

[0141] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the battery health state calculation method as described in the embodiments.

[0142] This application also provides a vehicle equipped with a battery health status calculation device as described in the embodiment, the battery health status calculation device executing the battery health status calculation method described in the embodiment.

[0143] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0146] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0147] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0148] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0149] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0150] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for calculating battery health status, characterized in that, The method includes: Acquire historical discharge data of the battery under test, including battery temperature, discharge factor, depth of discharge, and number of depths of discharge in different usage stages; the depth of discharge is represented by a gradient, which is determined based on the remaining capacity of the battery under test at the last moment of each usage stage; The historical discharge data of the battery under test is preprocessed. The preprocessed historical discharge data is input into the state calculation model to obtain the current battery health status of the battery under test. The state calculation model includes a first calculation network and a second calculation network jointly trained based on a historical sample dataset. The historical sample dataset includes battery temperature samples, discharge rate samples, discharge depth samples, and depth counts of multiple battery samples at different usage stages. The first calculation network identifies the degradation degree of the battery samples, and the second calculation network combines the degradation degree identification result output by the first calculation network with the historical sample dataset to output the health status identification result of each of the multiple battery samples. The gradient is determined based on the remaining capacity of the battery under test at the last moment of each usage phase, including: Obtain the remaining capacity of the battery under test at the last moment of each usage phase; Among multiple preset remaining capacity ranges, the remaining capacity range to which the remaining capacity belongs is determined, and the gradient corresponding to the remaining capacity range is used as the gradient of the depth of discharge of the battery under test in this usage stage. The multiple preset remaining capacity intervals are intervals obtained by dividing the remaining capacity values ​​according to preset intervals, and each remaining capacity interval corresponds to a different gradient.

2. The method according to claim 1, characterized in that, The state calculation model is trained according to the following steps: The historical sample dataset is constructed by acquiring historical sample data of multiple battery samples at different usage stages and the current real battery health status of each of the multiple battery samples. The historical sample data includes: battery temperature samples, discharge rate samples, discharge depth samples and depth counts. The historical sample dataset is preprocessed; Construct the first computing network and the second computing network respectively; The first computing network, taking the preprocessed historical sample dataset as input, identifies the degradation degree of each battery sample among the multiple battery samples; The second computing network takes the preprocessed historical sample dataset and the identification result of the degradation degree of each battery sample by the first computing network as input, and outputs the identification result of the battery health status of each battery sample among the multiple battery samples; Based on the battery health status identification results of each battery sample output by the second computing network, the parameters of the first computing network and the second computing network are updated.

3. The method according to claim 2, characterized in that, Both the first and second computing networks include an input layer, a hidden layer, and an output layer, wherein... The transfer function from the input layer to the hidden layer is: ， The transfer function from the hidden layer to the output layer is: 。 4. The method according to claim 3, characterized in that, The formula for calculating the j-th neuron in the hidden layer is: In the formula, The weights from the i-th neuron in the input layer to the j-th neuron in the hidden layer are denoted as . The value of the i-th neuron in the input layer. is the threshold of the j-th neuron in the hidden layer, and I is the total number of neurons in the input layer; From the hidden layer to the output layer, the formula for calculating the k-th neuron in the output layer is: In the formula, M j The value of the j-th neuron in the hidden layer. The weights from the j-th neuron in the hidden layer to the k-th neuron in the output layer are denoted as . J is the threshold of the k-th neuron in the output layer, and J is the total number of neurons in the hidden layer.

5. The method according to claim 4, characterized in that, Based on the battery health status identification results of each battery sample output by the second computing network, the parameters of the first computing network and the second computing network are updated, including: Based on the identification result of the battery health status of each battery sample output by the second computing network, the error between the identification result and the current true battery health status of each battery sample is determined. When the error is greater than or equal to the calibration value, the parameters of the first computing network and the second computing network are iteratively updated. When the error is less than the calibration value, the weights and thresholds from the input layer to the hidden layer of the first computing network and the weights and thresholds from the hidden layer to the output layer of the second computing network are used as the model parameters of the trained state computing model.

6. A battery health status calculation device, characterized in that, The device includes: The data acquisition module is used to acquire historical discharge data of the battery under test. The historical discharge data includes the battery temperature, discharge factor, discharge depth, and number of depths of the battery under test in different usage stages. The discharge depth is represented by a gradient, which is determined based on the remaining capacity of the battery under test at the last moment of each usage stage. A preprocessing module is used to preprocess the historical discharge data of the acquired battery under test; The state calculation module is used to input the preprocessed historical discharge data into the state calculation model to obtain the current battery health state of the battery under test. The state calculation model includes a first calculation network and a second calculation network jointly trained based on a historical sample dataset. The historical sample dataset includes battery temperature samples, discharge rate samples, discharge depth samples, and depth counts of multiple battery samples at different usage stages. The first calculation network identifies the degradation degree of the battery samples, and the second calculation network combines the degradation degree identification result output by the first calculation network with the historical sample dataset to output the health status identification result of each of the multiple battery samples. The gradient is determined based on the remaining capacity of the battery under test at the last moment of each usage phase, including: Obtain the remaining capacity of the battery under test at the last moment of each usage phase; Among multiple preset remaining capacity ranges, the remaining capacity range to which the remaining capacity belongs is determined, and the gradient corresponding to the remaining capacity range is used as the gradient of the depth of discharge of the battery under test in this usage stage. The multiple preset remaining capacity intervals are intervals obtained by dividing the remaining capacity values ​​according to preset intervals, and each remaining capacity interval corresponds to a different gradient.

7. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the battery health state calculation method as described in any one of claims 1 to 5.

8. A vehicle, characterized in that, The vehicle is equipped with a battery health status calculation device as described in claim 6, and the battery health status calculation device performs the battery health status calculation method as described in any one of claims 1 to 5.

Images