Energy storage battery aging prediction method and device and storage medium

Abstract

The invention discloses an energy storage battery aging prediction method and device and a storage medium, and belongs to the technical field of battery management, and the method comprises the steps: obtaining the state data of an energy storage battery in an aging period, and the aging period comprises a complete charging process or a complete discharging process; determining a characteristic parameter according to the state data, determining an attention weight based on a self-attention mechanism according to the characteristic parameter, and inputting the characteristic parameter and the attention weight into a battery aging prediction model; and determining an aging prediction result of the energy storage battery according to an output result of the battery aging prediction model. The invention aims to improve the accuracy of predicting the aging degree of the energy storage battery and improve the safety of the energy storage battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to a method, device and storage medium for predicting aging of an energy storage battery. Background Art

[0002] Energy storage batteries are widely used in various fields due to their advantages such as long life cycle, high capacity, high energy density and high specific energy. However, as their operating time increases, the internal part of the energy storage battery will undergo irreversible aging due to changes in electrochemical composition, which may eventually lead to thermal runaway failures and safety accidents. Therefore, there is a need to regularly predict and evaluate the aging degree of energy storage batteries. At present, the aging prediction method for energy storage batteries is generally based on the analysis of various historical operating data of energy storage batteries to predict the aging degree of energy storage batteries. However, the prediction accuracy of this prediction method depends on whether it can select the appropriate type of analysis data, which is easy to lead to insufficient prediction accuracy and inability to accurately judge the aging degree of energy storage batteries, affecting the safety of energy storage batteries. Summary of the invention

[0003] The present invention provides an energy storage battery aging prediction method, device and storage medium to solve one or more technical problems present in the prior art, and at least provide a beneficial choice or create conditions.

[0004] The present invention provides a method for predicting aging of an energy storage battery, the method comprising: Acquiring status data of the energy storage battery during an aging cycle, wherein the aging cycle includes a complete charging process or a complete discharging process; The characteristic parameters are determined based on the state data, the attention weight based on the self-attention mechanism is determined according to the characteristic parameters, and the characteristic parameters and the attention weight are input into the battery aging prediction model; The aging prediction result of the energy storage battery is determined according to the output result of the battery aging prediction model.

[0005] Optionally, the working mode of the energy storage battery includes a constant current mode or a constant voltage mode, and the step of determining the characteristic parameter according to the state data includes: Determine a first voltage change rate according to voltage data and discharge time of the energy storage battery discharged in the constant current mode; Determine a second voltage change rate according to voltage data and charging time of the energy storage battery charged in the constant current mode; Determining a target charging time according to the charging time of the energy storage battery in the constant current mode; Determine the current change rate according to the current data and charging time of the energy storage battery charged in the constant voltage mode; Determining a temperature change rate according to temperature data and charging time of the energy storage battery when charged in the constant voltage mode; Wherein, the characteristic parameter includes at least one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate.

[0006] Optionally, the step of determining the attention weight based on the self-attention mechanism according to the feature parameter includes: Determining data elements of the self-attention mechanism according to the feature parameter, the number of the data elements being greater than one; The first type of feature vector corresponding to each of the data elements is determined based on the data element, the preset weight vector and the first linear transformation parameter, and the second type of feature vector corresponding to each of the data elements is determined based on the data element, the preset weight vector and the second linear transformation parameter; Taking one of the first type of eigenvectors and one of the second type of eigenvectors as a group, traversing all possible combinations, calculating the quantity product of the first type of eigenvectors and the second type of eigenvectors in each combination, and obtaining the corresponding correlation vector for each combination; The association vector is processed by a Softmax function to obtain the attention weight.

[0007] Optionally, the step of determining the data element of the self-attention mechanism according to the feature parameter includes: When the characteristic parameters include more than one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate, and the temperature change rate, the data element includes the corresponding first voltage change rate, the second voltage change rate, the target charging time, the current change rate, and / or the temperature change rate.

[0008] Optionally, the step of determining the data element of the self-attention mechanism according to the feature parameter includes: When the characteristic parameters include one of data of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, the aging period is divided into a plurality of sub-periods according to the preset time step, the sub-parameters corresponding to the characteristic parameters in each of the sub-periods are determined, and the sub-parameters corresponding to the sub-periods in each of the sub-periods are determined, and the sub-parameters are determined as the data element.

[0009] Optionally, the battery aging prediction model is a DNN model, the DNN model is provided with four dense layers, the activation function of the dense layer is ReLU, and a Dropout layer is provided between the hidden layer and the output layer of the battery aging prediction model.

[0010] Optionally, the energy storage battery aging prediction method further includes: When the battery aging prediction model is trained, the prediction error value of the battery aging prediction model is evaluated by the root mean square error, and the prediction error percentage of the battery aging prediction model is evaluated by the average absolute percentage error; When the prediction error value is less than or equal to the preset error value, and the prediction error percentage is less than or equal to the preset error percentage, it is determined that the battery aging prediction model training is completed.

[0011] Optionally, the aging prediction result includes a state data curve of the energy storage battery at different aging degrees within the aging cycle.

[0012] In addition, in order to achieve the above-mentioned object, the present application also proposes an energy storage battery aging prediction device, which includes: a memory, a processor, and an energy storage battery aging prediction program stored on the memory and can be run on the processor, and the energy storage battery aging prediction program is formulated as a step of implementing the energy storage battery aging prediction method as described in any one of the above.

[0013] In addition, in order to achieve the above-mentioned object, the present application also proposes a storage medium on which an energy storage battery aging prediction program is stored. When the energy storage battery aging prediction program is executed by a processor, the steps of implementing the energy storage battery aging prediction method as described in any one of the above.

[0014] The present invention has at least the following beneficial effects: by acquiring the state data of the energy storage battery in the aging cycle, the operating status of the energy storage battery in the complete charging process or the complete discharging process is understood, and the characteristic parameters are determined according to the state data. Before using the battery aging prediction model to analyze the characteristic parameters, the attention weight is determined according to the characteristic parameters based on the self-attention mechanism, and the characteristic parameters and the attention weight are input into the battery aging prediction model to realize the prediction, and the aging prediction result of the energy storage battery is determined from the output result of the battery aging prediction model. Compared with the analysis of various historical operation data of the energy storage battery separately, the present invention introduces the self-attention mechanism, and the connection between the characteristic parameters is mined through the self-attention mechanism, and the corresponding attention weight is formed, so that the battery aging prediction model has more comprehensive data support, improves its prediction accuracy, realizes accurate judgment of the aging degree of the energy storage battery, and improves the safety of the energy storage battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings are used to provide a further understanding of the technical solution of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the technical solution of the present invention and do not constitute a limitation on the technical solution of the present invention.

[0016] Figure 1It is a flow chart of the first embodiment of the energy storage battery aging prediction method of the present invention; Figure 2 This is an example diagram of a state data change curve of the energy storage battery aging prediction method of the present invention; Figure 3 A schematic diagram of the calculation flow of the self-attention mechanism of the energy storage battery aging prediction method of the present invention; Figure 4 This is a schematic diagram of the device structure involved in the operation of an embodiment of an energy storage battery aging prediction device of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0018] It should be noted that, although the functional modules are divided in the system schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0019] The embodiment of the present invention provides a method for predicting energy storage battery aging. Figure 1 , a first embodiment of the energy storage battery aging prediction method of the present application is proposed. In this embodiment, the energy storage battery aging prediction method includes: Step S10, obtaining status data of the energy storage battery during an aging cycle, where the aging cycle includes a complete charging process or a complete discharging process.

[0020] In this embodiment, it is defined that the energy storage battery undergoes a complete charging process or a complete discharging process as undergoing an aging cycle. Since the complete charging process or the complete discharging process can intuitively reflect the change in the storage capacity of the energy storage battery, in order to improve the prediction accuracy of the aging degree of the energy storage battery, this embodiment selects to analyze based on the complete charging process or the complete discharging process. On the other hand, the state data of the energy storage battery in the aging cycle is the state data of the energy storage battery in a complete charging process or a complete discharging process. The state data includes the current data, voltage data, temperature and power value of the energy storage battery, and the power value can reflect the two data of charging time or discharging time.

[0021] Step S20, determining characteristic parameters according to the state data, determining attention weights based on the self-attention mechanism according to the characteristic parameters, and inputting the characteristic parameters and attention weights into the battery aging prediction model.

[0022] Since the present application is based on the analysis of the aging cycle, the state data obtained at each moment in the aging cycle is integrated into feature parameters for analysis. At the same time, in order to improve the prediction accuracy, the self-attention mechanism is introduced. The self-attention mechanism can realize the understanding of the sequence by calculating the association between each element in the sequence, calculate the correlation between the elements, and assign different attention levels to each element. Therefore, this embodiment calculates the feature parameters through the self-attention mechanism to obtain the corresponding attention weights, forms a data sequence with the feature parameters and attention weights, and inputs the data sequence into the battery aging prediction model for prediction.

[0023] Step S30, determining the aging prediction result of the energy storage battery according to the output result of the battery aging prediction model.

[0024] The aging cycle of this embodiment can also be used to quantify the aging degree of the energy storage battery. For example, if the aging degree of battery A is 30 aging cycles, it means that the aging degree of battery A is equivalent to 30 complete charging or discharging processes. It should be noted that this does not represent the number of incomplete charging or discharging times actually experienced by battery A. Based on this, the aging prediction result determined by the output result of the battery aging prediction model in this embodiment can evaluate how many aging cycles the current aging degree of the energy storage battery is equivalent to. At the same time, it can also predict the operating status of the energy storage battery in more aging cycles, thereby providing users with an intuitive and quantifiable aging degree reference.

[0025] The embodiment of the present invention has at least the following beneficial effects: by acquiring the state data of the energy storage battery in the aging cycle, the operating status of the energy storage battery in the complete charging process or the complete discharging process is understood, the characteristic parameters are determined according to the state data, and before the characteristic parameters are analyzed by the battery aging prediction model, the attention weight is determined according to the characteristic parameters based on the self-attention mechanism, and the characteristic parameters and the attention weight are input into the battery aging prediction model to realize the prediction, and the aging prediction result of the energy storage battery is determined from the output result of the battery aging prediction model. Compared with the analysis of each historical operation data of the energy storage battery separately, the present invention introduces the self-attention mechanism, and the connection between the characteristic parameters is mined through the self-attention mechanism, and the corresponding attention weight is formed, so that the battery aging prediction model has more comprehensive data support, improves its prediction accuracy, realizes accurate judgment of the aging degree of the energy storage battery, and improves the safety of the energy storage battery.

[0026] Furthermore, in this embodiment, the working mode of the energy storage battery includes a constant current mode or a constant voltage mode, and the step of determining the characteristic parameter according to the state data includes: The first voltage change rate is determined according to voltage data and discharge time of the energy storage battery discharged in the constant current mode.

[0027] The second voltage change rate is determined according to voltage data and charging time of the energy storage battery charged in the constant current mode.

[0028] The target charging time is determined according to the charging time of the energy storage battery in the constant current mode.

[0029] The current change rate is determined based on the current data and charging time of the energy storage battery charged in the constant voltage mode.

[0030] The temperature change rate is determined based on the temperature data and charging time of the energy storage battery when charged in the constant voltage mode.

[0031] The characteristic parameters include at least one of a first voltage change rate, a second voltage change rate, a target charging time, a current change rate and a temperature change rate.

[0032] The working modes of the energy storage battery include a constant current mode and a constant voltage mode. In the constant current mode, the discharge current of the battery remains constant, and the battery voltage will gradually decrease as the power decreases. In the constant voltage mode, the battery tries to maintain a constant voltage level during the discharge process, but the current will gradually decrease as the power decreases. Although the two modes are generally used to evaluate the characteristics of the energy storage battery during the discharge process, this embodiment also takes into account the charging process under the two modes to determine the characteristic parameters.

[0033] Specifically, in the constant current mode, the first voltage change rate is calculated according to the voltage data in a complete discharge process and the discharge time required for the entire discharge process, referring to Figure 2 , Figure 2 The voltage change curves of several energy storage batteries with different aging degrees under complete discharge in constant current mode are shown. The aging degree of the curves from the rightmost to the left gradually deepens. The rightmost curve is the curve at 30 aging cycles. Secondly, the second voltage change rate is calculated based on the voltage data in a complete charging process and the charging time required for the entire charging process, and the charging time required for the entire charging process is used as the target charging time; while in constant voltage mode, the current change rate is calculated based on the current data in a complete charging process and the charging time required for the entire charging process, and the temperature change rate is calculated based on the temperature data in a complete charging process and the charging time required for the entire charging process. The above-mentioned first voltage change rate, second voltage change rate, target charging time, current change rate and temperature change rate are used as five indicators for predicting the aging degree of energy storage batteries in this embodiment. The selection of indicators is selected by technicians through comparison with other types of indicators and prediction experiments, and the indicators with the best prediction accuracy are selected.

[0034] It should be noted that the characteristic parameters include at least one of the above five indicators, that is, the user can choose to use only any of them or use more than one indicator according to his actual needs. All the five indicators can separately realize the function of predicting the aging degree of energy storage battery.

[0035] By combining the two working modes of the energy storage battery constant current mode and the constant voltage mode, as well as data in the charging process and discharge process, the characteristic parameters can cover multiple indicators and improve the freedom of users to select indicators. At the same time, more accurate prediction can be achieved based on the characteristic parameters composed of five indicators and improve prediction accuracy.

[0036] Furthermore, in this embodiment, the step of determining the attention weight based on the self-attention mechanism according to the feature parameters includes: The data elements of the self-attention mechanism are determined according to the feature parameters, and the number of the data elements is greater than one.

[0037] The first type of feature vector corresponding to each data element is determined based on the data element, the preset weight vector and the first linear transformation parameter, and the second type of feature vector corresponding to each data element is determined based on the data element, the preset weight vector and the second linear transformation parameter.

[0038] Taking a first type of eigenvector and a second type of eigenvector as a group, traversing all possible combinations, calculating the quantity product of the first type of eigenvector and the second type of eigenvector in each combination, and obtaining the corresponding correlation vector for each combination.

[0039] The associated vector is processed by the Softmax function to obtain the attention weight.

[0040] The step of determining the data elements of the self-attention mechanism according to the feature parameters comprises: When the characteristic parameters include more than one of a first voltage change rate, a second voltage change rate, a target charging time, a current change rate, and a temperature change rate, the data element includes a corresponding first voltage change rate, a second voltage change rate, a target charging time, a current change rate, and / or a temperature change rate.

[0041] In this embodiment, the characteristic parameters include more than one data of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, that is, more than one indicator. For easy understanding, the self-attention mechanism is explained by the characteristic parameters including five indicators. Specific, reference Figure 3 , Figure 3 For the calculation process diagram of the self-attention mechanism of this embodiment, when the characteristic parameters include five indicators, the number of data elements corresponds to the number of types of feature parameters, that is, the five indicators can be directly used as five different data elements, in Figure 3A1, A2, A3 and A4 are used as illustrations respectively, the first linear transformation parameter is parameter q, the second linear transformation parameter is parameter k, the preset weight vector, parameter q and parameter k are all parameters and vectors pre-set by technicians after comparison through control variable experiments, the data element is weighted with the preset weight vector and the first linear transformation parameter to obtain the first type of feature vector, the data element is weighted with the preset weight vector and the second type of linear transformation parameter to obtain the second type of feature vector, taking data element A1 as an example, A1 is weighted with the preset weight vector and parameter q to obtain feature vector q1, A1 is weighted with the preset weight vector and parameter w to obtain feature vector k1, then a first type feature vector and a second type feature vector are taken as a group, all possible combinations are traversed, and the scalar product of the first type feature vector and the second type feature vector in the combination is taken as the association vector of the combination, Figure 3 The example of combining vectors q1 and k1 is used to calculate the vector , vector q1 is combined with k2 to obtain the vector , and so on, for the obtained association vector, it is processed by the Softmax function to obtain the corresponding attention weight, for example Figure 3 In and wait.

[0042] By taking the feature parameters as the data elements of the self-attention mechanism and calculating their attention weights, the battery aging prediction model can combine the correlation between several indicators for analysis, improve the prediction accuracy, accurately judge the aging degree of the energy storage battery, and improve the safety of the energy storage battery.

[0043] Furthermore, in this embodiment, the battery aging prediction model is a DNN model, the DNN model is provided with 4 dense layers, the activation function of the dense layer is ReLU, and a Dropout layer is provided between the hidden layer and the output layer of the battery aging prediction model.

[0044] Specifically, the battery aging prediction model of this embodiment is a prediction model built and trained based on a deep neural network (DNN). The model includes 4 dense layers and uses ReLU as an activation function. Using ReLU as an activation function makes the model easier to train and usually achieves better performance.

[0045] A Dropout layer is set between the hidden layer and the output layer of the battery aging prediction model. The Dropout layer assists the model in regularization. While combining the above-mentioned self-attention mechanism to improve the prediction accuracy, it also reduces the risk of overfitting of the model.

[0046] In addition, the battery aging prediction model also combines Adam optimization to compile the results to further improve the prediction accuracy of the battery aging prediction model.

[0047] Furthermore, in this embodiment, the energy storage battery aging prediction method further includes: When the battery aging prediction model is trained, the prediction error value of the battery aging prediction model is evaluated by the root mean square error, and the prediction error percentage of the battery aging prediction model is evaluated by the mean absolute percentage error.

[0048] When the prediction error value is less than or equal to the preset error value, and the prediction error percentage is less than or equal to the preset error percentage, it is determined that the battery aging prediction model training is completed.

[0049] Specifically, the root mean square error (RMSE) is calculated as:

[0050] The calculation formula for mean absolute percentage error (MAPE) is:

[0051] in, is the actual value, is the predicted value output by the model during training, and n is the number of predicted values.

[0052] The root mean square error is used to measure the actual numerical difference between the predicted value and the actual value, and the mean absolute percentage error is used to measure the error between the predicted value and the actual value in the form of a percentage. When training the battery aging prediction model, the error calculated by the root mean square error calculation formula is defined as the prediction error value, and the error calculated by the mean absolute percentage error calculation formula is defined as the prediction error percentage. Combining the root mean square error and the mean absolute percentage error, the battery aging prediction model training is determined to be completed only when both errors meet the requirements, that is, the prediction error value is less than or equal to the preset error value, and the prediction error percentage is less than or equal to the preset error percentage.

[0053] By combining the two errors to evaluate the performance of the battery aging prediction model during training, the trained model has higher accuracy requirements, thereby improving the prediction accuracy of the battery aging prediction model.

[0054] Furthermore, in this embodiment, the aging prediction result includes a state data curve of the energy storage battery at different aging degrees within the aging cycle.

[0055] In this embodiment, the aging prediction result determined according to the output result of the battery aging prediction model can evaluate how many aging cycles the current aging degree of the energy storage battery is equivalent to, and at the same time, it can also predict the operating status of the energy storage battery in more aging cycles.

[0056] First, for the collected status data of the energy storage battery, you can refer to Figure 2 The voltage change curve is displayed in the form of. If the collected state data also includes other types of data such as current, they can also be displayed separately. This embodiment only takes voltage as an example to illustrate that the voltage change curve formed by the collected state data is compared with the preset Figure 2 By comparing similar graphs, we can infer how many aging cycles the current aging degree of the energy storage battery is equivalent to. Taking the inference that the current aging degree of the energy storage battery is equivalent to 30 aging cycles as an example, according to the output results of the battery aging prediction model, we can further obtain the aging cycles of the energy storage battery at 120 aging cycles, 210 aging cycles and 300 aging cycles (refer to Figure 2 ) The voltage change curve, current change curve or temperature change curve at different aging degrees are similar. It should be noted that the output result of the battery aging prediction model corresponds to the type of characteristic parameter input. For example, when the input characteristic parameter does not include the temperature change rate, the corresponding output result does not include temperature-related data. At this time, the aging prediction result does not need to display temperature-related data.

[0057] Through the status data curves of different aging degrees formed in the aging prediction results, users can intuitively understand the current operating status of the energy storage battery, as well as the operating status when the aging degree is more serious, thereby gaining an intuitive understanding of how the performance of the energy storage battery changes with the aging degree, and improving the experience of energy storage battery aging prediction.

[0058] Further, based on the above embodiment, a second embodiment of the energy storage battery aging prediction method of the present application is proposed. In this embodiment, the step of determining the data element of the self-attention mechanism according to the characteristic parameters includes: When the characteristic parameter includes one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, the aging period is divided into multiple sub-periods according to a preset time step, and the sub-parameters corresponding to the characteristic parameter in each sub-period are determined, and the sub-parameters are determined as data elements.

[0059] Specifically, when the characteristic parameter includes only one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, that is, only one indicator is used, in order to meet the requirements of multiple data elements required for the calculation of the self-attention mechanism, a single type of characteristic parameter can be converted into multiple sub-parameters, wherein the aging cycle is first divided into multiple sub-cycles according to the preset time step, and each sub-cycle is regarded as a single aging cycle in the first embodiment, and the corresponding parameters are calculated according to the state data in each sub-cycle, so as to obtain the sub-parameters, and each sub-parameter is used as a data element; for example, the aging cycle is 30 minutes, the preset time step is 5 minutes, and the calculation of the first voltage change rate is taken as an example. In the first embodiment, the first voltage change rate within 30 minutes is calculated, and only one value is obtained. In this embodiment, the 6 first voltage change rates within 6 sub-cycles divided by the preset time step are calculated, which can be determined as the corresponding 6 data elements. Further, for the obtained multiple data elements, the remaining steps can refer to the first embodiment, that is, the corresponding attention weight can be obtained based on one indicator.

[0060] By using a preset time step to divide the aging cycle into multiple sub-cycles, when the characteristic parameters include only one indicator, the self-attention mechanism can also calculate the corresponding attention weight, so that the aging degree of the energy storage battery can be predicted through one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate. The correlation between the indicators at different stages of the aging cycle can also be explored, thereby improving the versatility of the energy storage battery aging prediction method of the present application.

[0061] In addition, during the training process of the battery aging prediction model, in order to enable the model to realize the prediction function and maintain satisfactory prediction accuracy regardless of the input of multiple characteristic parameters or the input of multiple sub-parameters converted from a single characteristic parameter, training can also be performed based on the input of multiple sub-parameters converted from a single characteristic parameter, and the root mean square error and mean absolute percentage error are also combined to evaluate the prediction accuracy of each indicator when it is input separately, thereby improving the versatility of the battery aging prediction model.

[0062] In other embodiments, when the characteristic parameter includes more than one type of data among the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, it can also be combined with the step of dividing the aging cycle into multiple sub-cycles according to the preset time step, and determining the sub-parameters corresponding to the characteristic parameters in each sub-cycle, so that each type of characteristic parameter can also obtain multiple corresponding data elements, which is conducive to the analysis of the self-attention mechanism and improves the prediction accuracy of the battery aging prediction model.

[0063] On the other hand, reference Figure 4 , Figure 4 It is a schematic diagram of the device structure of the energy storage battery aging prediction device.

[0064] An energy storage battery aging prediction device is provided, comprising: a processor and a memory; wherein the memory is used to store a computer-readable program. When the computer-readable program is executed by the processor, the processor implements the energy storage battery aging prediction method as described in any one of the above technical solutions.

[0065] It will be appreciated by those skilled in the art that all or some of the steps and systems in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. As is well known to those skilled in the art, communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0066] In addition, an embodiment of the present invention further provides a storage medium on which an energy storage battery aging prediction program is stored. When the energy storage battery aging prediction program is executed by a processor, the relevant steps of any embodiment of the above energy storage battery aging prediction method are implemented.

[0067] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0068] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0069] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0070] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0071] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0072] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0073] Although the description of the present application has been quite detailed and specifically describes several embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, but should be regarded as providing a broad possible interpretation of these claims by reference to the attached claims, taking into account the prior art, so as to effectively cover the intended scope of the present application. In addition, the above description of the present application is based on the embodiments foreseeable by the inventor, and its purpose is to provide a useful description, and those non-substantial changes to the present application that have not yet been foreseen may still represent equivalent changes to the present application.

Claims

1. A method for predicting energy storage battery aging, characterized in that: The energy storage battery aging prediction method comprises: Acquiring status data of the energy storage battery during an aging cycle, wherein the aging cycle includes a complete charging process or a complete discharging process; Determine characteristic parameters according to the state data, determine attention weights based on a self-attention mechanism according to the characteristic parameters, and input the characteristic parameters and the attention weights into a battery aging prediction model; The aging prediction result of the energy storage battery is determined according to the output result of the battery aging prediction model.

2. The energy storage battery aging prediction method according to claim 1, characterized in that: The working mode of the energy storage battery includes a constant current mode or a constant voltage mode, and the step of determining the characteristic parameter according to the state data includes: Determine a first voltage change rate according to voltage data and discharge time of the energy storage battery discharged in the constant current mode; Determine a second voltage change rate according to voltage data and charging time of the energy storage battery charged in the constant current mode; Determining a target charging time according to the charging time of the energy storage battery in the constant current mode; Determine the current change rate according to the current data and charging time of the energy storage battery charged in the constant voltage mode; Determining a temperature change rate according to temperature data and charging time of the energy storage battery when charged in the constant voltage mode; Wherein, the characteristic parameter includes at least one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate.

3. The energy storage battery aging prediction method according to claim 2, characterized in that: The step of determining the attention weight based on the self-attention mechanism according to the feature parameter comprises: Determining data elements of the self-attention mechanism according to the feature parameter, the number of the data elements being greater than one; Determine a first type feature vector corresponding to each of the data elements according to the data elements, the preset weight vector and the first linear transformation parameter, and determine a second type feature vector corresponding to each of the data elements according to the data elements, the preset weight vector and the second linear transformation parameter; Taking one of the first-category feature vectors and one of the second-category feature vectors as a group, traverse all possible combinations, calculate the dot product of the first-category feature vector and the second-category feature vector in each combination, and obtain an association vector corresponding to each combination; The association vector is processed by a Softmax function to obtain the attention weight.

4. The energy storage battery aging prediction method according to claim 3, characterized in that: The step of determining the data element of the self-attention mechanism according to the feature parameter comprises: When the characteristic parameter includes more than one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, the data element includes the corresponding first voltage change rate, the second voltage change rate, the target charging time, the current change rate and / or the temperature change rate.

5. The energy storage battery aging prediction method according to claim 3, characterized in that: The step of determining the data element of the self-attention mechanism according to the feature parameter comprises: When the characteristic parameter includes one of the first voltage change rate, the second voltage change rate, the target charging time, the current change rate and the temperature change rate, the aging cycle is divided into multiple sub-cycles according to a preset time step, and the sub-parameters corresponding to the characteristic parameters in each of the sub-cycles are determined, and the sub-parameters are determined as the data elements.

6. The energy storage battery aging prediction method according to claim 1, characterized in that: The battery aging prediction model is a DNN model, which is provided with 4 dense layers, the activation function of the dense layer is ReLU, and a Dropout layer is provided between the hidden layer and the output layer of the battery aging prediction model.

7. The energy storage battery aging prediction method according to claim 1, characterized in that: The energy storage battery aging prediction method also includes: When the battery aging prediction model is trained, the prediction error value of the battery aging prediction model is evaluated by the root mean square error, and the prediction error percentage of the battery aging prediction model is evaluated by the mean absolute percentage error; When the prediction error value is less than or equal to the preset error value, and the prediction error percentage is less than or equal to the preset error percentage, it is determined that the battery aging prediction model training is completed.

8. The energy storage battery aging prediction method according to any one of claims 1 to 7, characterized in that: The aging prediction result includes a state data curve of the energy storage battery at different aging degrees within the aging cycle.

9. An energy storage battery aging prediction device, characterized in that: The energy storage battery aging prediction device includes: a memory, a processor, and an energy storage battery aging prediction program stored in the memory and executable on the processor, wherein the energy storage battery aging prediction program is configured to implement the steps of the energy storage battery aging prediction method according to any one of claims 1 to 8.

10. A storage medium, characterized in that: The storage medium stores an energy storage battery aging prediction program, and when the energy storage battery aging prediction program is executed by the processor, the steps of the energy storage battery aging prediction method according to any one of claims 1 to 8 are implemented.

Images