A lithium-ion battery health status assessment and prediction method and system

Abstract

The present invention belongs to the technical field of battery life prediction and discloses a lithium-ion battery health status assessment and prediction method and system, including: obtaining training data and test data representing the discharge state of the lithium-ion battery, preprocessing the data collected by the sensor to obtain the training data and test data representing the discharge state of the lithium-ion battery; constructing a slope feature data set based on the lithium-ion battery discharge curve, obtaining a feature data set that affects the health state of the lithium-ion battery, and constructing a battery health indicator SOH; constructing and training a battery health status assessment model based on the K-Means++ clustering method and a battery remaining service life prediction model based on the GRU neural network, constructing and training the battery health status assessment model based on the K-Means++ clustering method and the battery remaining service life prediction model based on the GRU neural network, and using the trained model to assess the health status of the lithium-ion battery and predict the remaining service life.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery life prediction, and in particular relates to a method and system for evaluating and predicting the health status of a lithium-ion battery. Background Art

[0002] Currently, the closest existing technologies for predicting battery remaining useful life primarily employ model-based and data-driven approaches. Model-based approaches predict the degradation state and trajectory by learning the battery's internal physicochemical reactions or establishing a battery failure model. These approaches include electrochemical models and equivalent circuit models, often combined with Kalman filters and particle filters. However, the complex physicochemical processes during battery charging and discharging make it difficult to obtain a suitable physical model to describe the system's dynamic characteristics. Therefore, model-based approaches are not ideal in practical applications. Data-driven approaches are categorized as artificial intelligence (AI) and statistical data-driven. These methods utilize degradation data and do not require precise analysis of the degradation process. Instead, they employ various data analysis methods (such as machine learning, stochastic processes, and statistical signal processing) to identify the degradation information implicit in the data and predict the battery's remaining useful life. However, these methods require less degradation data and are therefore more suitable for assessing battery health. However, to improve model performance, they must be combined with optimization algorithms.

[0003] At the current level of technology, predicting the health status of lithium-ion batteries usually ignores the intrinsic connections and mutual influences between different sensor signals. Therefore, directly extracting features may not effectively reflect the health status of lithium-ion batteries, and training models based on these features may reduce the generalization performance of the model.

[0004] Currently, battery health status assessment and prediction methods are usually limited to predicting the remaining service life of the battery, ignoring the comprehensive assessment of the health status, and therefore cannot accurately locate the health status of lithium-ion batteries.

[0005] In summary, existing technologies have limitations in predicting the remaining useful life of lithium-ion batteries. This is because the directly extracted feature information is limited, resulting in poor generalization performance of the prediction model. Furthermore, simply predicting the remaining useful life of a battery fails to accurately monitor its health status.

[0006] The difficulty of solving the above technical problems: Consider the intrinsic connection and mutual influence between signals, and extract the slope features to evaluate the health status of the battery and classify it.

[0007] The significance of solving the above technical problems: The applicability of existing technologies is limited. The present invention aims to propose a method for evaluating and predicting the health status of lithium-ion batteries based on the slope characteristics of the discharge curve. Summary of the Invention

[0008] In response to the problems existing in the prior art, the present invention provides a method and system for evaluating and predicting the health status of lithium-ion batteries.

[0009] The present invention is implemented as follows: a method for evaluating and predicting the health status of a lithium-ion battery, comprising:

[0010] The first step is to preprocess the data collected by the sensor, including denoising, missing value and outlier processing;

[0011] In the second step, a slope feature dataset is constructed based on the lithium-ion battery discharge curve to obtain a feature dataset that affects the lithium-ion battery state of health (SOH). A regression model is used to construct the health indicator SOH.

[0012] The third step is to build a battery health status assessment model based on the K-Means++ clustering method;

[0013] The fourth step is to build a battery remaining service life prediction model based on the GRU neural network;

[0014] Step 5: Divide the dataset into training data and test data. Use the training data to train the battery health status assessment model and the battery remaining service life prediction model. Then use the test data to test the above models.

[0015] The sixth step is to use the K-Means++ clustering model and the GRU prediction model to evaluate and predict the health status of the lithium-ion battery and stack the results. First, the battery health level is determined, and then its remaining service life is determined.

[0016] Furthermore, the second step of constructing a slope feature using the lithium-ion battery discharge curve includes the following steps:

[0017] (1) Obtain the four signals of the terminal voltage, output current, temperature and time used for each cycle during lithium-ion battery discharge, and extract the initial discharge voltage U D0 , discharge valley voltage U DMIN , discharge valley voltage time t DMIN , initial discharge temperature T D0 , discharge peak temperature T DMAX , discharge peak temperature time t DMAX , discharge valley current I DMIN , discharge peak current I DMAX , discharge valley current time t DMIN , discharge peak current time t DMAX These 10 characteristics;

[0018] (2) Using the above 10 features, construct slope features, including the discharge voltage slope feature K U , discharge current slope characteristic K I , discharge temperature slope characteristic K T , the specific formula is as follows:

[0019]

[0020]

[0021]

[0022] (3) Establish a regression model, including input layer, GRU layer, fully connected layer, and regression layer. In this model, the input layer unit represents the slope feature K U , K I , K T ,The regression layer unit is the health index SOH. The regression model is trained to obtain the health index SOH.

[0023] Furthermore, the method of training the battery health status assessment model and the battery remaining service life prediction model in the fifth step includes:

[0024] (1) Establish a K-Means++ clustering model, determine the cluster types and cluster centers, and use the training data to train the K-Means++ clustering model;

[0025] (2) Establish a GRU prediction model, which includes a GRU neural network with an input layer, a GRU layer, a fully connected layer, and an output layer. In this model, the input layer unit represents the health indicator SOH, and the output layer unit represents the prediction result. The GRU prediction model is trained using training data.

[0026] (3) Use the test data to test the K-Means++ clustering model and GRU prediction model.

[0027] Furthermore, the sixth step of using the trained K-Means++ clustering model and GRU neural network model to evaluate the health status and predict the remaining service life of the battery includes: given sensor data, extracting the slope feature data set, and constructing the health indicator SOH, and then using the K-Means++ clustering model and GRU prediction model to realize the evaluation and prediction of the battery health status.

[0028] Another object of the present invention is to provide a lithium-ion battery health status assessment and prediction system for implementing the lithium-ion battery health status assessment and prediction method, comprising:

[0029] Health indicator SOH construction module: used to analyze the sensor signals collected during lithium-ion battery discharge, extract slope features, and use the feature data to construct the health indicator SOH;

[0030] Health status assessment model construction module: used to input the unlabeled health indicator SOH into the K-Means++ clustering model, train the K-Means++ clustering model, and build a battery health status assessment model;

[0031] Battery health status assessment module: used to use the battery health status assessment model to perform unsupervised clustering of the unlabeled health indicators (SOH) of the battery to be assessed, thereby achieving the assessment of the health status of lithium-ion batteries;

[0032] Life prediction model building module: used to input the health indicator SOH into the GRU neural network model, train the GRU neural network model, and build a battery remaining service life prediction model;

[0033] Battery Remaining Service Life Prediction Module: It is used to input the health indicator SOH of the battery to be predicted into the trained GRU neural network model to realize the remaining service life prediction of the lithium-ion battery.

[0034] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the lithium-ion battery health status assessment and prediction method.

[0035] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the lithium-ion battery health status assessment and prediction method.

[0036] Another object of the present invention is to provide an information data processing terminal, which is used to implement the lithium-ion battery health status assessment and prediction system.

[0037] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0038] First, in view of the technical problems existing in the above-mentioned prior art and the difficulty of solving these problems, we closely combine the technical solutions to be protected by the present invention and the results and data during the research and development process, and conduct a detailed and in-depth analysis of how the technical solutions of the present invention solve the technical problems and some creative technical effects brought about by solving the problems. The specific description is as follows:

[0039] This method extracts slope features from four signals during battery discharge: terminal voltage, output current, temperature, and cycle time. These features are then fed into a GRU neural network for regression to produce the state of health (SOH) indicator. The extraction of these new features takes into account the inherent connections and interactions between these signals, encompassing more information about lithium-ion battery discharge. This significantly improves the model's generalization performance and broadens its application.

[0040] This invention first assesses the health status of a lithium-ion battery and then predicts its remaining useful life. However, existing methods for assessing and predicting the health status of lithium-ion batteries often directly predict the remaining useful life of the battery, ignoring the importance of assessing its health status. Therefore, using this method for assessing and predicting the health status of a battery can achieve higher accuracy. After modification, this method can be widely applied to assessing the health status and predicting the remaining useful life of various lithium-ion batteries.

[0041] Second, considering the technical solution as a whole or from the perspective of the product, the technical effects and advantages of the technical solution to be protected by the present invention are described in detail as follows:

[0042] The method for evaluating and predicting the health status of lithium-ion batteries based on the slope characteristics of the discharge curve proposed in the present invention utilizes K-Means++ clustering and GRU neural networks, giving full play to the advantages of each. At the same time, it takes into account the data characteristics of different sensor signals during the discharge process of the lithium-ion battery, deeply mines the information of the sensor data, and improves the generalization performance of the model. The method is simple to operate and can be widely applied to various battery predictions. Compared with directly predicting the remaining service life of lithium-ion batteries, this method has higher accuracy in evaluating and predicting the health status of batteries. In the field of lithium-ion battery health status prediction, the method for evaluating and predicting the health status of lithium-ion batteries based on the slope characteristics of the discharge curve proposed in the present invention is highly innovative and practical.

[0043] Third, as auxiliary evidence for the inventiveness of the claims of the present invention, it is also reflected in the following important aspects:

[0044] (1) The expected benefits and commercial value of the technical solution of the present invention after transformation are:

[0045] By accurately assessing the health status of lithium-ion batteries and predicting their remaining useful life, this invention enables more effective battery lifecycle management, extending battery life, reducing replacement and maintenance costs, and improving device reliability and stability. In the energy storage and electric transportation sectors, this invention can reduce unnecessary battery replacements, improve system efficiency, and thus reduce overall costs.

[0046] (2) Whether the technical solution of the present invention solves the technical problems that people have been eager to solve but have not been able to solve successfully:

[0047] The technical solution of this invention utilizes a more detailed and comprehensive feature extraction method, taking into account the inherent connections and interactions between different sensor signals during lithium-ion battery discharge. It also combines clustering methods with neural network models to improve the model's generalization and accuracy. This comprehensive approach facilitates more precise assessment of battery health status and enables more accurate prediction of remaining battery life. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 This is a flow chart of a method for evaluating and predicting the health status of a lithium-ion battery based on the discharge curve slope characteristics provided by an embodiment of the present invention.

[0050] Figure 2 This is a flow chart for implementing a method for evaluating and predicting the health status of a lithium-ion battery based on the discharge curve slope characteristics provided by an embodiment of the present invention.

[0051] Figure 3 It is a schematic diagram of slope characteristics provided by the implementation of the present invention.

[0052] Figure 4 2 is a schematic diagram of the results obtained by using the K-Means++ clustering model provided in an embodiment of the present invention to evaluate the battery health status and the GRU neural network to predict the remaining battery service life; wherein (a) is the result obtained using the K-Means++ clustering model; (b) is the result obtained using the GRU neural network model.

[0053] Figure 5 This is a structural diagram of a lithium-ion battery health status assessment and prediction system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] In response to the problems existing in the prior art, the present invention provides a method and system for evaluating and predicting the health status of a lithium-ion battery. The present invention is described in detail below with reference to the accompanying drawings.

[0056] 1. Explanatory Examples In order to enable those skilled in the art to fully understand how to implement the present invention, this section provides an illustrative example that expands upon the technical solutions of the claims.

[0057] like Figure 1 As shown, the method for evaluating and predicting the health status of a lithium-ion battery based on the discharge curve slope characteristics provided by an embodiment of the present invention includes the following steps:

[0058] S101: Preprocessing the data collected by the sensor;

[0059] S102: Analyze the battery discharge signal, extract the slope feature, and use the feature data to construct the health indicator SOH;

[0060] S103: Build a battery health status assessment model based on K-Means++;

[0061] S104: Constructing a GRU-based battery remaining service life prediction model;

[0062] S105: Divide the data set into training data and test data;

[0063] S106: Using the training data, train the K-Means++ clustering model and the GRU neural network model;

[0064] S107: Use the test data to test the K-Means++ clustering model and the GRU neural network model;

[0065] S108: Input the battery health indicator SOH into the K-Means++ clustering model and the GRU neural network model to evaluate and predict the health status of the lithium-ion battery.

[0066] In a preferred embodiment of the present invention, the lithium-ion battery is replaced or repaired according to the battery health status and remaining service life.

[0067] In a preferred embodiment of the present invention, step S102 extracts slope features based on the lithium-ion battery discharge curve and constructs the health indicator SOH as follows:

[0068] Step 1: Obtain the four signals of the lithium-ion battery during discharge: terminal voltage, output current, temperature, and time used for each cycle, and extract the initial discharge voltage U D0 , discharge valley voltage U DMIN , discharge valley voltage time tDMIN , initial discharge temperature T D0 , discharge peak temperature T DMAX , discharge peak temperature time t DMAX , discharge valley current I DMIN , discharge peak current I DMAX , discharge valley current time t DMIN , discharge peak current time t DMAX These 10 characteristics;

[0069] Step 2: Using the above 10 features, construct slope features, including discharge voltage slope feature K U , discharge current slope characteristic K I , discharge temperature slope characteristic K T , the specific formula is as follows:

[0070]

[0071]

[0072]

[0073] Step 3: Establish a regression model, including input layer, GRU layer, fully connected layer, and regression layer. In this model, the input layer unit represents the slope feature K U , K I , K T ,The regression layer unit is the health index SOH. The regression model is trained to obtain the health index SOH.

[0074] In a preferred embodiment of the present invention, the K-Means++ model construction method of step S103 is: establishing a K-Means++ clustering model, determining cluster types and cluster centers, and training the K-Means++ clustering model using training data.

[0075] In a preferred embodiment of the present invention, the GRU neural network model construction method of step S104 is: establishing a GRU prediction model, including a GRU neural network of an input layer, a GRU layer, a fully connected layer, and an output layer. In this model, the input layer unit represents the health indicator SOH, and the output layer unit represents the prediction result. The GRU prediction model is trained using training data.

[0076] The technical effects of the present invention are described in detail below in conjunction with tests.

[0077] The data used in this embodiment comes from lithium-ion battery aging data published by NASA's Ames Research Center. This dataset represents operational experiments of lithium-ion batteries during their charge and discharge processes. To study the health of lithium-ion batteries, different types of sensors were used to collect data. The dataset includes various battery charge and discharge scenarios, recording different sensor data from the start of charge and discharge to the end of life.

[0078] Data from different battery life cycles were selected to divide the training set and test set. The battery health status assessment and remaining service life prediction were performed using the steps described in the present invention, the K-Means++ clustering model, and the GRU neural network model. The results are as follows: Figure 4 (a) is the result obtained using the K-Means++ clustering model. Figure 4 (b) is the result obtained using the GRU neural network model.

[0079] According to the results, it can be seen that the battery health status assessment and prediction method proposed in the present invention is closer to the actual value of battery life loss than the result obtained by directly predicting the remaining service life, and can better reflect the remaining service life of the battery.

[0080] like Figure 5 As shown, the lithium-ion battery health status assessment and prediction system provided by the embodiment of the present invention includes:

[0081] Health indicator SOH construction module: used to analyze the sensor signals collected during lithium-ion battery discharge, extract slope features, and use the feature data to construct the health indicator SOH;

[0082] Health status assessment model construction module: used to input the unlabeled health indicator SOH into the K-Means++ clustering model, train the K-Means++ clustering model, and build a battery health status assessment model;

[0083] Battery health status assessment module: used to use the battery health status assessment model to perform unsupervised clustering of the unlabeled health indicators (SOH) of the battery to be assessed, thereby achieving the assessment of the health status of the lithium-ion battery;

[0084] Life prediction model building module: used to input the health indicator SOH into the GRU neural network model, train the GRU neural network model, and build a battery remaining service life prediction model;

[0085] Battery Remaining Service Life Prediction Module: It is used to input the health indicator SOH of the battery to be predicted into the trained GRU neural network model to realize the remaining service life prediction of the lithium-ion battery.

[0086] An application embodiment of the present invention provides a computer device, which includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor executes the steps of the lithium-ion battery health status assessment and prediction method.

[0087] An application embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of a method for evaluating and predicting the health status of a lithium-ion battery.

[0088] An application embodiment of the present invention provides an information data processing terminal, which is used to implement a lithium-ion battery health status assessment and prediction system.

[0089] 2. Application Examples: In order to demonstrate the creativity and technical value of the technical solution of the present invention, this section provides application examples of the claimed technical solution on specific products or related technologies.

[0090] The application examples of the present invention utilize the B0005, B0006, and B0007 lithium-ion battery aging data released by NASA Ames Research Center. First, the aging data for different batteries is preprocessed, and then slope features are extracted to construct the state of health (SOH) indicator. The health indicator data is divided into training data and test data. The training data is used to train a battery state of health assessment model and a remaining useful life (RUS) prediction model. Next, the test data is used to validate both models. Finally, the battery state of health assessment and prediction results are stacked.

[0091] 3. Evidence of the effects of the embodiments: The embodiments of the present invention have achieved some positive effects during the development or use process, and indeed have great advantages over the existing technology. The following content describes them with reference to the data, charts, etc. of the experimental process.

[0092] The application embodiment of the present invention achieves good results in feature extraction. The three slope features K U , K I , K T The correlations with health indicators SOH were all greater than 0.96, among which K I , K T The correlation with SOH reached 0.99.

[0093] The application embodiment of the present invention achieves good results in the evaluation and prediction of battery health status. Figure 4 (a) shows the health status assessment results of B0006 battery, Figure 4 (b) shows the remaining service life prediction results of B0006 battery.

[0094] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware portion can be implemented using dedicated logic; the software portion can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art will appreciate that the above-mentioned devices and methods can be implemented using computer-executable instructions and / or contained in processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, or can be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0095] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for evaluating and predicting the health status of a lithium-ion battery, characterized in that: include: The first step is to preprocess the data collected by the sensor, including denoising, missing value and outlier processing; In the second step, a slope feature dataset is constructed based on the lithium-ion battery discharge curve to obtain a feature dataset that affects the lithium-ion battery state of health (SOH). A regression model is used to construct the health indicator SOH. The third step is to build a battery health status assessment model based on the K-Means++ clustering method; The fourth step is to build a battery remaining service life prediction model based on the GRU neural network; Step 5: Divide the dataset into training data and test data. Use the training data to train the battery health status assessment model and the battery remaining service life prediction model. Then use the test data to test the above models. In the sixth step, the K-Means++ clustering model and the GRU prediction model are used to evaluate and predict the health status of the lithium-ion battery and stack the results. First, the battery health status level is determined, and then its remaining service life is determined. The second step of constructing a slope feature using the lithium-ion battery discharge curve includes the following steps: (1) Obtain the four signals of the terminal voltage, output current, temperature and time used for each cycle during lithium-ion battery discharge, and extract the initial discharge voltage U D0 , discharge valley voltage U DMIN , discharge valley voltage time t DMIN , initial discharge temperature T D0 , discharge peak temperature T DMAX , discharge peak temperature time t DMAX , discharge valley current I DMIN , discharge peak current I DMAX , discharge valley current time t DMIN , discharge peak current time t DMAX These 10 characteristics; (2) Using the above 10 features, construct slope features, including the discharge voltage slope feature K U , discharge current slope characteristic K I , discharge temperature slope characteristic K T , the specific formula is as follows: (3) Establish a regression model, including input layer, GRU layer, fully connected layer, and regression layer. In this model, the input layer unit represents the slope feature K U , K I , K T ,The regression layer unit is the health index SOH. The regression model is trained to obtain the health index SOH.

2. The lithium-ion battery health status assessment and prediction method according to claim 1, wherein: The method of training the battery health status assessment model and the battery remaining service life prediction model in the fifth step includes: (1) Establish a K-Means++ clustering model, determine the cluster types and cluster centers, and use the training data to train the K-Means++ clustering model; (2) Establish a GRU prediction model, which includes a GRU neural network with an input layer, a GRU layer, a fully connected layer, and an output layer. In this model, the input layer unit represents the health indicator SOH, and the output layer unit represents the prediction result. The GRU prediction model is trained using training data. (3) Use the test data to test the K-Means++ clustering model and GRU prediction model.

3. The lithium-ion battery health status assessment and prediction method according to claim 1, wherein: The sixth step of using the trained K-Means++ clustering model and GRU neural network model to evaluate the health status and predict the remaining service life of the battery includes: given sensor data, extracting the slope feature data set, and constructing the health indicator SOH, and then using the K-Means++ clustering model and GRU prediction model to evaluate and predict the battery health status.

4. A lithium-ion battery health status assessment and prediction system implementing the lithium-ion battery health status assessment and prediction method according to any one of claims 1 to 3, characterized in that: include: Health indicator SOH construction module: used to analyze the sensor signals collected during lithium-ion battery discharge, extract slope features, and use the feature data to construct the health indicator SOH; Health status assessment model construction module: used to input the unlabeled health indicator SOH into the K-Means++ clustering model, train the K-Means++ clustering model, and build a battery health status assessment model; Battery health status assessment module: used to use the battery health status assessment model to perform unsupervised clustering of the unlabeled health indicators (SOH) of the battery to be assessed, thereby achieving the assessment of the health status of lithium-ion batteries; Life prediction model building module: used to input the health indicator SOH into the GRU neural network model, train the GRU neural network model, and build a battery remaining service life prediction model; Battery Remaining Service Life Prediction Module: It is used to input the health indicator SOH of the battery to be predicted into the trained GRU neural network model to realize the remaining service life prediction of the lithium-ion battery.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the lithium-ion battery health status assessment and prediction method according to any one of claims 1 to 3.

6. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the lithium-ion battery health status assessment and prediction method according to any one of claims 1 to 3.

7. An information data processing terminal, the information data processing terminal being used to implement the lithium-ion battery health status assessment and prediction system as claimed in claim 4.

Images