WO2024066055A1

WO2024066055A1 - Service life prediction and state estimation method for bidirectional lithium ion battery

Info

Publication number: WO2024066055A1
Application number: PCT/CN2022/138134
Authority: WO
Inventors: 郭媛君; 安钊; 杨之乐; 吴承科; 胡天宇
Original assignee: 深圳先进技术研究院
Priority date: 2022-09-30
Filing date: 2022-12-09
Publication date: 2024-04-04
Also published as: CN115267558A

Abstract

A service life prediction and state estimation method for a bidirectional lithium ion battery, comprising: acquiring battery characteristic data corresponding to each moment of a target bidirectional lithium ion battery in a continuous time period; according to the battery characteristic data corresponding to each moment, determining time sequence data corresponding to the target bidirectional lithium ion battery; inputting the time sequence data into a target neural network model to obtain a first prediction state corresponding to the target bidirectional lithium ion battery; inputting the battery characteristic data corresponding to each moment into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium ion battery; and according to the first prediction state and the second prediction state, determining a predicted service life corresponding to the target bidirectional lithium ion battery.

Description

A method for life prediction and state estimation of bidirectional lithium-ion batteries

Technical Field

The present invention relates to the field of data processing, and in particular to a life prediction and state estimation method for a bidirectional lithium-ion battery.

Background technique

Traditional battery life prediction is mainly based on manual experience analysis, which requires a large number of relevant professionals to measure the battery voltage, current, pressure difference, temperature, etc. on site, and then judge the remaining life of the battery based on experience. This not only has certain dangerous factors, but also requires a lot of manpower costs. More importantly, due to the existence of subjectivity, it is easy to lead to inaccurate judgment results.

Therefore, the existing technology still needs to be improved and developed.

technical problem

The technical problem to be solved by the present invention is that, in view of the above-mentioned defects of the prior art, a method for life prediction and state estimation of a bidirectional lithium-ion battery is provided, aiming to solve the problem that the prior art determines the battery life based on manual experience analysis, which requires a lot of labor costs and leads to inaccurate judgment results due to subjective reasons.

Technical Solutions

The technical solution adopted by the present invention to solve the problem is as follows:

In a first aspect, an embodiment of the present invention provides a method for life prediction and state estimation of a bidirectional lithium-ion battery, wherein the method comprises:

Obtain battery characteristic data corresponding to a target bidirectional lithium-ion battery at a plurality of times in a continuous time period;

Determine the time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to the plurality of moments;

Inputting the time series data into a target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery;

Inputting the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium-ion battery;

The predicted lifespan of the target bidirectional lithium-ion battery is determined according to the first predicted state and the second predicted state.

In one embodiment, the time series data is a time series graph, the horizontal axis of the time series graph is time, and the vertical axis is amplitude; the time series graph includes a plurality of data points, and the plurality of data points correspond one-to-one to a plurality of moments, the horizontal axis corresponding to each of the data points is determined based on the moment corresponding to the data point, and the vertical axis corresponding to each of the data points is determined based on the battery characteristic data at the moment corresponding to the data point.

In one embodiment, the target neural network model includes a feature extraction module and a classification module, and the time series data is input into the target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery, including:

Inputting the time series graph into the feature extraction module to obtain a polar coordinate graph corresponding to the time series graph, wherein the polar coordinate graph includes a plurality of polar coordinate points, and the plurality of polar coordinate points correspond to the plurality of data points one by one;

The polar coordinate graph is input into the classification module to obtain the first predicted state.

In one implementation, the step of inputting the time series graph into the feature extraction module to obtain a polar coordinate graph corresponding to the time series graph includes:

Input the time series graph into the feature extraction module, and determine the polar coordinate point sets corresponding to the data points respectively through the feature extraction module, wherein each polar coordinate point set includes a plurality of polar coordinate points, and the polar coordinate points in the same polar coordinate point set are in a rotationally symmetric relationship;

The polar coordinate graph is determined according to polar coordinate point sets corresponding to the data points.

In one implementation, a method for determining the polar coordinate point set corresponding to each of the data points includes:

Obtaining the maximum amplitude and the minimum amplitude corresponding to the time series graph;

Obtaining a preset target value, and determining the associated data point corresponding to each of the data points according to the target value;

Acquire a plurality of preset angles, wherein the plurality of angles are sequentially increased or decreased by a preset angle value;

The phases of the polar coordinate points corresponding to the data point are determined according to the angles, the amplitudes corresponding to the associated data points, the maximum amplitude and the minimum amplitude, wherein the phases corresponding to the polar coordinate points correspond to the angles one-to-one.

In one embodiment, determining the phases of the plurality of polar coordinate points corresponding to the data point according to the plurality of angles, the amplitudes corresponding to the associated data points, the maximum amplitude value, and the minimum amplitude value includes:

Inputting the amplitudes, the maximum amplitudes and the minimum amplitudes corresponding to the associated data points in sequence into a target calculation formula, and obtaining the phases corresponding to the polar coordinate points in the polar coordinate point set corresponding to the data point;

The target calculation formula is:

in,

is the amplitude,

is the phase,

is the maximum amplitude value,

is the minimum amplitude value, t is the target value,

is one of the several angles, and g is a preset fixed value.

In one embodiment, the predicted life is a target remaining number of charge and discharge times, and determining the predicted life corresponding to the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state includes:

Determining a first remaining number of charge and discharge times corresponding to the target bidirectional lithium-ion battery according to the first predicted state;

Determining a second remaining number of charge and discharge times corresponding to the target bidirectional lithium-ion battery according to the second predicted state;

The target remaining charge and discharge times is determined according to the first remaining charge and discharge times and the second remaining charge and discharge times.

In a second aspect, an embodiment of the present invention further provides a life prediction and state estimation device for a bidirectional lithium-ion battery, wherein the device comprises:

A data acquisition module, used to acquire battery characteristic data corresponding to a target bidirectional lithium-ion battery at a plurality of moments in a continuous time period;

A first prediction module, used for inputting the time series data into a target neural network model to obtain a first prediction state corresponding to the target bidirectional lithium-ion battery;

A second prediction module, used to input the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium-ion battery;

A life prediction module is used to determine the predicted life of the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state.

In a third aspect, an embodiment of the present invention further provides a terminal, wherein the terminal includes a memory and one or more processors; the memory stores one or more programs; the program includes instructions for executing the life prediction and state estimation method of a bidirectional lithium-ion battery as described in any of the above; and the processor is used to execute the program.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a plurality of instructions stored thereon, wherein the instructions are suitable for being loaded and executed by a processor to implement the steps of any of the above-mentioned methods for life prediction and state estimation of a bidirectional lithium-ion battery.

Beneficial Effects

Beneficial effects of the present invention: The embodiment of the present invention obtains the battery characteristic data corresponding to each moment of the target bidirectional lithium-ion battery in a continuous time period; determines the time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to each moment; inputs the time series data into the target neural network model to obtain the first predicted state corresponding to the target bidirectional lithium-ion battery; inputs the battery characteristic data corresponding to each moment into the target deep learning prediction model to obtain the second predicted state corresponding to the target bidirectional lithium-ion battery; determines the predicted life corresponding to the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state. The present invention predicts the battery life of a bidirectional lithium-ion battery through a neural network model and a deep learning model, solving the problem in the prior art of determining the battery life based on manual experience analysis, which has high labor costs and inaccurate judgment results.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

FIG1 is a schematic flow chart of a method for life prediction and state estimation of a bidirectional lithium-ion battery provided in an embodiment of the present invention.

[Corrected 12.04.2023 in accordance with Article 91]
FIG. 2 is a time series diagram provided by an embodiment of the present invention. FIG. 3 is a polar coordinate diagram provided by an embodiment of the present invention.

[Corrected 12.04.2023 in accordance with Article 91]
FIG. 4 is a schematic diagram of the internal modules of the device for life prediction and state estimation of a bidirectional lithium-ion battery provided in an embodiment of the present invention.

[Corrected 12.04.2023 in accordance with Article 91]
FIG5 is a functional block diagram of a terminal provided by an embodiment of the present invention.

Embodiments of the present invention

The present invention discloses a life prediction and state estimation method for a bidirectional lithium-ion battery. In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. 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.

It will be understood by those skilled in the art that, unless otherwise stated, the singular forms "one", "the", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and/or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and/or" used herein includes all or any unit and all combinations of one or more associated listed items.

It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as herein.

In view of the above-mentioned defects of the prior art, the present invention provides a method for life prediction and state estimation of a bidirectional lithium-ion battery, the method obtaining battery characteristic data corresponding to a target bidirectional lithium-ion battery at several moments in a continuous time period; determining the time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to the several moments; inputting the time series data into a target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery; inputting the battery characteristic data corresponding to the several moments into a target deep learning prediction model to obtain a second predicted state corresponding to the target bidirectional lithium-ion battery; and determining the predicted life corresponding to the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state. The present invention predicts the battery life of a bidirectional lithium-ion battery by combining a neural network model and a deep learning model, thereby solving the problem in the prior art that the battery life is determined based on manual experience analysis, which requires a lot of labor costs and leads to inaccurate judgment results due to subjective reasons.

As shown in FIG1 , the method comprises the following steps:

Step S100: Obtain battery characteristic data corresponding to a target bidirectional lithium-ion battery at a plurality of moments in a continuous time period.

Specifically, the target bidirectional lithium-ion battery in this embodiment can be any battery that currently needs to be predicted for battery life. Since the battery characteristic data can reflect the current battery state of the target bidirectional lithium-ion battery, such as temperature, voltage, current, etc., this embodiment needs to obtain the battery characteristic data of the target bidirectional lithium-ion battery at several moments in a continuous time period, and determine the battery life of the target bidirectional lithium-ion battery by analyzing the changes in the battery characteristic data at these moments.

As shown in FIG1 , the method further comprises the following steps:

Step S200: determining time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to the plurality of moments.

Specifically, since this embodiment needs to use a neural network model to predict battery life later, and the neural network model has a fixed input format, this embodiment needs to convert the acquired battery characteristic data at each moment into a time series format that can be processed by the neural network model, that is, to obtain time series data.

In one implementation, the time series data is a time series graph, the horizontal axis of the time series graph is time, and the vertical axis is amplitude; the time series graph includes a plurality of data points, and the plurality of data points correspond one-to-one to a plurality of moments, the horizontal axis corresponding to each of the data points is determined based on the moment corresponding to the data point, and the vertical axis corresponding to each of the data points is determined based on the battery characteristic data at the moment corresponding to the data point.

Specifically, as shown in FIG2 , the time series data in this embodiment exists in the form of a time series graph. There are multiple data points in the time series graph, and each data point represents the battery characteristic data at a moment. For each data point, its horizontal coordinate is determined based on the corresponding moment, and its vertical coordinate is determined based on the corresponding battery characteristic data. By analyzing the position distribution of each data point in the time series graph, the current state of the target bidirectional lithium-ion battery can be obtained.

As shown in FIG1 , the method further comprises the following steps:

Step S300: input the time series data into a target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery.

Specifically, the present embodiment pre-trains a target neural network model. Since the target neural network model has been pre-trained with a large amount of training data and has learned the data characteristics of time series data of different battery states, the currently obtained time series data is input into the target neural network model. The target neural network model can classify the target bidirectional lithium-ion battery based on the input time series data, thereby outputting the first predicted state currently corresponding to the target bidirectional lithium-ion battery.

In one implementation, the target neural network model includes a feature extraction module and a classification module, and step S300 specifically includes the following steps:

Step S301, inputting the time series graph into the feature extraction module to obtain a polar coordinate graph corresponding to the time series graph, wherein the polar coordinate graph includes a plurality of polar coordinate points, and the plurality of polar coordinate points correspond to the plurality of data points one by one;

Step S302: input the polar coordinate diagram into the classification module to obtain the first predicted state.

Specifically, the target neural network model in this embodiment mainly includes two parts, one is a feature extraction module, and the other is a classification module. For the feature extraction module, after the time series graph is input into the feature extraction module, the feature extraction module will determine the corresponding polar coordinate point for each data point, and output a polar coordinate graph containing the polar coordinate points corresponding to each data point. For the classification module, the polar coordinate graph is input into the classification module, and the classification module will classify the current state of the target bidirectional lithium-ion battery according to the position distribution of each polar coordinate in the input polar coordinate graph, and then output the first predicted state corresponding to the target bidirectional lithium-ion battery. As shown in Figure 2, since the positions of the data points in the time series graph are closely distributed, it is difficult to extract the slight changes in the position distribution of each data point based on the original time series graph. As shown in Figure 3, this embodiment converts the time series graph into a polar coordinate graph first, which can amplify the slight changes in the position distribution of each data point, thereby prompting the classification module to obtain more accurate classification results.

In one implementation, step S301 specifically includes the following steps:

Step S3011, inputting the time series graph into the feature extraction module, and determining the polar coordinate point sets corresponding to the data points respectively through the feature extraction module, wherein each polar coordinate point set includes a plurality of polar coordinate points, and the polar coordinate points in the same polar coordinate point set are in a rotationally symmetric relationship.

Specifically, in the present embodiment, each data point corresponds to a plurality of equal polar coordinate points in the polar coordinate diagram. For each data point, the feature extraction module will first determine a base point in the polar coordinate diagram according to the time corresponding to the data point and the battery data feature, and then rotate the base point a preset number of times to obtain a plurality of sub-points corresponding to the base point, and form a polar coordinate point set corresponding to the data point based on the base point and the plurality of sub-points. Therefore, for each data point, there are a plurality of corresponding polar coordinate points for the data point, and each polar coordinate point is in a rotationally symmetric relationship. For example, as shown in FIG3 , the polar coordinate diagram presents a rotationally symmetric petal shape. The present embodiment increases the number of polar coordinate points in the polar coordinate diagram by generating a plurality of polar coordinate points corresponding to each data point, thereby further amplifying the slight changes in the position distribution of each data point, and prompting the classification module to obtain more accurate classification results.

Step S30111, obtaining the maximum amplitude and the minimum amplitude corresponding to the time series graph;

Step S30112, obtaining a preset target value, and determining the associated data point corresponding to each of the data points according to the target value;

Step S30113, obtaining a plurality of preset angles, wherein the plurality of angles are sequentially increased or decreased by a preset angle value;

Step S30114, determining the phases of the polar coordinate points corresponding to the data point according to the angles, the amplitudes corresponding to the associated data points, the maximum amplitude and the minimum amplitude, wherein the phases corresponding to the polar coordinate points correspond to the angles one-to-one.

In simple terms, each data point needs to determine its corresponding polar coordinate point set based on its corresponding associated data point, where the associated data point of each data point is separated from the data point by the sequence position of the target value. For example, the target value is t, and the associated data point corresponding to the data point _Xi is _Xi+t . Specifically, it is first necessary to obtain the amplitude extreme value of the time series graph, the amplitude of the associated data point Xi+ _t in the time series graph, and a plurality of pre-set angle values, and then determine the phase corresponding to each angle value based on the amplitude extreme value, the amplitude of the associated data point _Xi+t in the time series graph, and each angle value, and generate a polar coordinate point corresponding to the data point _Xi based on each phase.

In one implementation, step S30114 specifically includes the following steps:

The target calculation formula is:

in,

is the amplitude,

is the phase,

is the maximum amplitude value,

is the minimum amplitude value, t is the target value,

is one of the several angles, and g is a preset fixed value.

For example, suppose that several angles include , t=10, then the time series diagram shown in FIG2 will be transformed into the polar coordinate diagram shown in FIG3, presenting a rotationally symmetrical petal shape.

As shown in FIG1 , the method further comprises the following steps:

Step S400: input the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium-ion battery.

Simply put, the target neural network model is a machine learning model under deep supervised learning, while the target deep learning prediction model is a machine learning model under unsupervised learning. This embodiment uses two different types of machine learning models to jointly predict the battery state of the target bidirectional lithium-ion battery, which can avoid the prediction bias caused by a single model and improve the accuracy of battery state prediction. Specifically, this embodiment also pre-trains a target deep learning model, and inputs the battery feature data at each moment into the target deep learning prediction model to obtain the second predicted state of the target bidirectional lithium-ion battery.

In one implementation, the target neural network model and the target deep learning prediction model are respectively pre-trained with a target training data set, and the target training data set includes an original training data set and an augmented training data set, and the augmented training data set is obtained according to the original training data set and the DCGAN deep convolution generative adversarial network. This embodiment uses DCGAN to perform data expansion on the polar coordinate graph after feature extraction to obtain a large amount of data, thereby solving the problem of unbalanced and insufficient fault data. A target training data set containing sufficient data can increase the classification accuracy of the model and solve the problem of the robustness of the model.

In one implementation, the first predicted state/the second predicted state is one of an initial period, a healthy period, a decay period, and a scrap period, wherein the remaining number of charge and discharge times corresponding to the initial period, the healthy period, the decay period, and the scrap period decreases in sequence.

As shown in FIG1 , the method further comprises the following steps:

Step S500: Determine a predicted lifespan corresponding to the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state.

Specifically, since the first prediction state and the second prediction state are generated based on different types of machine learning models, respectively, in order to reduce the prediction deviation caused by a single model, this embodiment uses the first prediction state and the second prediction state to comprehensively determine the current battery state of the target bidirectional lithium-ion battery, thereby accurately predicting the corresponding predicted life of the target bidirectional lithium-ion battery.

In one implementation, the predicted life is a target remaining number of charge and discharge times, and step S500 specifically includes the following steps:

Step S501: determining a first remaining charge and discharge number corresponding to the target bidirectional lithium-ion battery according to the first predicted state;

Step S502: determining a second remaining number of charge and discharge times corresponding to the target bidirectional lithium-ion battery according to the second predicted state;

Step S503: Determine the target remaining charge and discharge times according to the first remaining charge and discharge times and the second remaining charge and discharge times.

In simple terms, this embodiment defines the remaining charge and discharge times of the target bidirectional lithium-ion battery as the target remaining charge and discharge times, and defines the battery life of the target bidirectional lithium-ion battery with the target remaining charge and discharge times. Specifically, since the first prediction state and the second prediction state are generated based on different types of machine learning models, respectively, in order to reduce the prediction deviation caused by a single model, this embodiment first determines the remaining charge and discharge times of the battery based on the first prediction state, that is, obtains the first remaining charge and discharge times, and then determines the remaining charge and discharge times of the battery with the second prediction state, that is, obtains the second remaining charge and discharge times. Finally, based on the first remaining charge and discharge times and the second remaining charge and discharge times, the target remaining charge and discharge times corresponding to the target bidirectional lithium-ion battery are comprehensively determined, that is, the predicted life of the target bidirectional lithium-ion battery is obtained.

In one implementation, step S503 specifically includes the following steps:

The target remaining charge and discharge times is determined according to an average value of the first remaining charge and discharge times and the second remaining charge and discharge times.

In another implementation, step S503 specifically includes the following steps:

Obtaining the model accuracy corresponding to the target neural network model and the target deep learning prediction model respectively;

Determine the weight values corresponding to the target neural network model and the target deep learning prediction model respectively according to the model accuracies corresponding to the target neural network model and the target deep learning prediction model respectively;

The target remaining charge and discharge times are obtained by taking a weighted average based on the weight values corresponding to the target neural network model and the target deep learning prediction model, the first remaining charge and discharge times, and the second remaining charge and discharge times.

In one implementation, the actual charging and discharging process of the target bidirectional lithium-ion battery can be simulated in the virtual world based on a data-driven model using digital twin technology, so as to obtain a number of battery characteristic data corresponding to the target bidirectional lithium-ion battery through real-time collection and monitoring.

In one implementation, the predicted life and status of the target bidirectional lithium-ion battery can be displayed in the virtual world of the digital twin, and the remaining number of charge and discharge cycles can be visualized, thereby enhancing the efficiency of human-computer interaction and the reliability of the digital twin system.

Based on the above embodiments, the present invention further provides a life prediction and state estimation device for a bidirectional lithium-ion battery, as shown in FIG4 , the device comprises:

The data acquisition module 01 is used to acquire the battery characteristic data corresponding to a plurality of moments in a continuous time period of the target bidirectional lithium-ion battery;

A first prediction module 02, used for inputting the time series data into a target neural network model to obtain a first prediction state corresponding to the target bidirectional lithium-ion battery;

The second prediction module 03 is used to input the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium-ion battery;

The life prediction module 04 is used to determine the predicted life of the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state.

Based on the above embodiments, the present invention also provides a terminal, whose principle block diagram can be shown in Figure 5. The terminal includes a processor, a memory, a network interface, and a display screen connected through a system bus. Among them, the processor of the terminal is used to provide computing and control capabilities. The memory of the terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the terminal is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a life prediction and state estimation method for a bidirectional lithium-ion battery is implemented. The display screen of the terminal can be a liquid crystal display screen or an electronic ink display screen.

Those skilled in the art will understand that the principle block diagram shown in FIG5 is only a block diagram of a partial structure related to the solution of the present invention, and does not constitute a limitation on the terminal to which the solution of the present invention is applied. The specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

In one implementation, one or more programs are stored in the memory of the terminal, and the terminal is configured to be executed by one or more processors. The one or more programs include instructions for performing a method for life prediction and state estimation of a bidirectional lithium-ion battery.

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and/or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

In summary, the present invention discloses a method for life prediction and state estimation of a bidirectional lithium-ion battery, the method obtaining battery characteristic data corresponding to a target bidirectional lithium-ion battery at a plurality of moments in a continuous time period; determining time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to the plurality of moments; inputting the time series data into a target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery; inputting the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second predicted state corresponding to the target bidirectional lithium-ion battery; and determining the predicted life corresponding to the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state. The present invention predicts the battery life of a bidirectional lithium-ion battery by combining a neural network model and a deep learning model, thereby solving the problem in the prior art that the battery life is determined based on manual experience analysis, which requires a large amount of labor costs and leads to inaccurate judgment results due to subjectivity.

It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

A method for life prediction and state estimation of a bidirectional lithium-ion battery, characterized in that the method comprises:

Obtain battery characteristic data corresponding to a target bidirectional lithium-ion battery at a plurality of times in a continuous time period;

Determine the time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to the plurality of moments;

Inputting the time series data into a target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery;

Inputting the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium-ion battery;

The predicted lifespan of the target bidirectional lithium-ion battery is determined according to the first predicted state and the second predicted state.
The life prediction and state estimation method of a bidirectional lithium-ion battery according to claim 1 is characterized in that the time series data is a time series graph, the abscissa of the time series graph is time, and the ordinate is amplitude; the time series graph includes a plurality of data points, and the plurality of data points correspond one-to-one to the plurality of moments, the abscissa corresponding to each of the data points is determined based on the moment corresponding to the data point, and the ordinate corresponding to each of the data points is determined based on the battery characteristic data at the moment corresponding to the data point.
The life prediction and state estimation method of a bidirectional lithium-ion battery according to claim 2 is characterized in that the target neural network model includes a feature extraction module and a classification module, and the time series data is input into the target neural network model to obtain a first predicted state corresponding to the target bidirectional lithium-ion battery, including:

Inputting the time series graph into the feature extraction module to obtain a polar coordinate graph corresponding to the time series graph, wherein the polar coordinate graph includes a plurality of polar coordinate points, and the plurality of polar coordinate points correspond to the plurality of data points one by one;

The polar coordinate graph is input into the classification module to obtain the first predicted state.
The life prediction and state estimation method of a bidirectional lithium-ion battery according to claim 3 is characterized in that the step of inputting the time series graph into the feature extraction module to obtain a polar coordinate graph corresponding to the time series graph comprises:

Input the time series graph into the feature extraction module, and determine the polar coordinate point sets corresponding to the data points respectively through the feature extraction module, wherein each polar coordinate point set includes a plurality of polar coordinate points, and the polar coordinate points in the same polar coordinate point set are in a rotationally symmetric relationship;

The polar coordinate graph is determined according to polar coordinate point sets corresponding to the data points.
The method for life prediction and state estimation of a bidirectional lithium-ion battery according to claim 1 is characterized in that a method for determining the polar coordinate point set corresponding to each of the data points comprises:

Obtaining the maximum amplitude and the minimum amplitude corresponding to the time series graph;

Obtaining a preset target value, and determining the associated data point corresponding to each of the data points according to the target value;

Acquire a plurality of preset angles, wherein the plurality of angles are sequentially increased or decreased by a preset angle value;

The phases of the polar coordinate points corresponding to the data point are determined according to the angles, the amplitudes corresponding to the associated data points, the maximum amplitude and the minimum amplitude, wherein the phases corresponding to the polar coordinate points correspond to the angles one-to-one.
The method for life prediction and state estimation of a bidirectional lithium-ion battery according to claim 5 is characterized in that the phase of the plurality of polar coordinate points corresponding to the data point is determined according to the plurality of angles, the amplitudes corresponding to the associated data points, the maximum amplitude value, and the minimum amplitude value, comprising:

Inputting the amplitudes, the maximum amplitudes and the minimum amplitudes corresponding to the associated data points in sequence into a target calculation formula, and obtaining the phases corresponding to the polar coordinate points in the polar coordinate point set corresponding to the data point;

The target calculation formula is:

in,
is the amplitude,
is the phase,
is the maximum amplitude value,
is the minimum amplitude value, t is the target value,
is one of the several angles, and g is a preset fixed value.
The method for life prediction and state estimation of a bidirectional lithium-ion battery according to claim 1, characterized in that the predicted life is a target remaining number of charge and discharge times, and determining the predicted life corresponding to the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state comprises:

Determining a first remaining number of charge and discharge times corresponding to the target bidirectional lithium-ion battery according to the first predicted state;

Determining a second remaining number of charge and discharge times corresponding to the target bidirectional lithium-ion battery according to the second predicted state;

The target remaining charge and discharge times is determined according to the first remaining charge and discharge times and the second remaining charge and discharge times.
A life prediction and state estimation device for a bidirectional lithium-ion battery, characterized in that the device comprises:

A data acquisition module, used to acquire battery characteristic data corresponding to a target bidirectional lithium-ion battery at a plurality of moments in a continuous time period;

Determine the time series data corresponding to the target bidirectional lithium-ion battery according to the battery characteristic data corresponding to the plurality of moments;

A first prediction module, used for inputting the time series data into a target neural network model to obtain a first prediction state corresponding to the target bidirectional lithium-ion battery;

A second prediction module, used to input the battery characteristic data corresponding to the plurality of moments into a target deep learning prediction model to obtain a second prediction state corresponding to the target bidirectional lithium-ion battery;

A life prediction module is used to determine the predicted life of the target bidirectional lithium-ion battery according to the first predicted state and the second predicted state.
A terminal, characterized in that the terminal includes a memory and one or more processors; the memory stores one or more programs; the program includes instructions for executing the life prediction and state estimation method of a bidirectional lithium-ion battery as described in any one of claims 1-7; and the processor is used to execute the program.
A computer-readable storage medium having a plurality of instructions stored thereon, characterized in that the instructions are suitable for being loaded and executed by a processor to implement the steps of the life prediction and state estimation method of a bidirectional lithium-ion battery as described in any one of claims 1 to 7 above.