Lithium ion battery online state monitoring and health state evaluation method and system

Abstract

The invention belongs to but is not limited to the technical field of battery health management and life prediction, and particularly relates to a lithium ion battery online state monitoring and health state evaluation method and system, and the method comprises the steps: collecting data collected by a sensor in the charging and discharging process of a lithium ion battery, carrying out the preprocessing, and extracting related features; and respectively constructing a battery state monitoring data set and a health state evaluation data set. A battery state monitoring data set is used and input into the random forest model to complete battery state monitoring, and meanwhile, the requirement for the real-time performance of the system is considered, and online monitoring of the battery state is achieved. And inputting the battery health state assessment data set into the GRU neural network to complete battery health state assessment. Characteristics used in the method are from the complete cycle process of the charging and discharging process of the lithium ion battery, more battery state change information can be obtained, and the accuracy of battery state monitoring and health state evaluation is improved.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the technical field of battery health management and life prediction, and in particular relates to a method and system for online status monitoring and health status assessment of a lithium-ion battery. Background Art

[0002] The battery management system can monitor and manage the battery. By calculating and collecting data such as battery voltage, current, and temperature, it can evaluate the battery state of charge (SOC) and state of health (SOH), and control the battery charging and discharging process, ultimately realizing battery state monitoring and health state evaluation, and reminding users to replace the battery before it reaches the end of its life. The existing methods mainly include direct measurement method, model-based method, and data-driven method. As the name suggests, the direct measurement method is the process of directly measuring parameters. However, in actual application, the working environment of lithium-ion batteries is complex and there are certain differences between different types of batteries. This method is more difficult to use. Model-based methods can be divided into methods based on physical models and chemical models. However, in actual application, this method requires prior knowledge based on physics and chemistry, and the complexity and versatility of the method remain to be verified. Data-driven methods have been widely used in recent years. This method does not require knowledge of electrochemistry and physics. It uses machine learning or deep learning methods based on certain historical data to obtain information from data, and ultimately achieves accurate SOH evaluation.

[0003] Problems with existing technologies: Existing technologies have certain limitations in lithium-ion battery status monitoring and health status assessment, including not considering the real-time requirements of battery status monitoring and not considering the battery degradation information contained in the full process characteristics. At the same time, it ignores the information brought by battery status monitoring to health status assessment, resulting in poor real-time performance of lithium-ion battery status monitoring and low accuracy of lithium-ion battery health status assessment results.

[0004] Difficulty in solving technical problems: Propose corresponding charging and discharging time characteristics, build an online platform for lithium-ion battery online status monitoring, and realize accurate and real-time lithium-ion battery online status monitoring and health status assessment.

[0005] Significance of solving technical problems: The feature extraction method and online status monitoring platform proposed in the present invention can overcome the limitations of the existing technology to a certain extent and improve the accuracy of lithium-ion battery health status assessment. Summary of the invention

[0006] In view of the problems existing in the prior art, the present invention combines artificial intelligence technology with new energy battery management, and provides a method and system for online state monitoring and health status assessment of lithium-ion batteries, which effectively solves the problem that the existing research on partial charging and discharging processes of lithium-ion batteries is difficult to fully reflect the state changes, and the technical bottleneck caused by the high real-time requirements for lithium-ion battery state monitoring. The method proposes the characteristics of the charging and discharging process, uses the random forest algorithm to realize accurate state monitoring of lithium-ion batteries, uses the digital signal processor (DSP) to respond to system abnormalities in a timely manner to realize online state monitoring, and also uses the gated recurrent unit (GRU) to realize accurate health status assessment of lithium-ion batteries, effectively assess the battery degradation state, improve the stability and reliability of the power system operation, and thus achieve maximum economic benefits. In addition, the method and system also have good adaptability and scalability, and can be widely used in new energy vehicles, power grid energy storage and other fields, providing strong technical support for the development of the industry.

[0007] The present invention is implemented as follows: a method for online status monitoring and health status assessment of a lithium-ion battery comprises the following steps:

[0008] Step 1, collecting data collected by the sensor during the charging and discharging process of the lithium-ion battery, and preprocessing the data;

[0009] Step 2: extract the battery features based on the data preprocessed in step 1, construct the battery health state SOH, and perform correlation analysis on the battery features, the status label and the battery health state SOH respectively;

[0010] Step 3, building a lithium-ion battery status monitoring model based on the random forest model;

[0011] Step 4, establishing a state monitoring data set according to the preprocessed data, battery characteristics and battery state labels, and dividing the state monitoring data set into a training set and a test set, training the lithium-ion battery state monitoring model according to the training set, and testing the lithium-ion battery state monitoring model according to the test set;

[0012] Step 5, evaluating the classification results of the lithium-ion battery state monitoring model according to multiple evaluation indicators to verify the accuracy of the lithium-ion battery state monitoring model;

[0013] Step 6, using a digital signal processor to construct an online state monitoring platform for lithium-ion batteries, and measuring the test time of the online state monitoring platform for lithium-ion batteries to verify the real-time performance of the online state monitoring platform;

[0014] Step 7, constructing a lithium-ion battery health status assessment model based on the gated recurrent unit neural network model;

[0015] Step 8, establishing a health status assessment data set based on the preprocessed data, battery characteristics, and battery health status SOH, and dividing the health status assessment data set into a training set and a test set, training a lithium-ion battery health status assessment model based on the training set, and testing the lithium-ion battery health status assessment model based on the test set;

[0016] Step 9: Evaluate the prediction results of the lithium-ion battery health status assessment model according to multiple evaluation indicators to verify the accuracy of the lithium-ion battery health status assessment model.

[0017] Further, the step 2 extracts the following steps for lithium-ion battery feature processing:

[0018] (1) Obtain the voltage and temperature data of the lithium-ion battery during the charging and discharging process, and obtain the duration of the battery constant current charging phase, the time when the temperature reaches the highest point during the charging phase, the time when the temperature reaches the mean value during the discharging phase, the duration of the charging phase, and the duration of the discharging phase.

[0019] (2) The obtained data is used to construct features based on the charging and discharging processes, including the proportion of constant current time during the charging process F1, the proportion of battery temperature reaching peak time during the charging process F2, and the proportion of battery temperature reaching mean time during the discharging process F3.

[0020] Get the battery capacity data and complete the construction of the battery health status, which is defined as follows:

[0021] Among them, C k represents the remaining available capacity of the battery after the kth cycle, C0 represents the rated capacity of the battery, and SOH k Indicates the battery health after the kth cycle.

[0022] (3) Constructing state monitoring and health status assessment data sets. According to the battery capacity data, the state is divided to obtain the state label. The battery health above 80% is normal and marked as 0. The health below 80% is abnormal and marked as 1. The extracted features and state labels are used to construct the state monitoring data set. The features extracted under normal state and the corresponding battery health state are used to construct the health status assessment data set.

[0023] The mutual information analysis method and the grey correlation analysis method are used to perform correlation analysis respectively to obtain the correlation degree between the feature and status labels and the battery capacity. The formula of the mutual information analysis method is as follows:

[0024] Among them, X is the proposed features F1, F2 and F3, and Y is the battery status label. The grey correlation analysis method is defined as follows:

[0025] Among them, ξ i (k) is feature F i The correlation coefficient of the kth cycle, r i F i The correlation degree with SOH, i = 1, 2, 3. ρ is the resolution coefficient, ρ∈(0, 1), the present invention takes ρ = 0.5, x i (k) is the proposed F i , N represents the sequence length (battery cycle period).

[0026] In the obtained analysis results, the results of the extracted features in the mutual information analysis were all above 0.4, with the highest being 0.66, and the result with the cycle period as the feature was 0.67. In the Pearson correlation coefficient analysis, the results were all above 0.99, while the results with the cycle period as the feature were all below 0.6. Taking all factors into consideration, the correlation of the proposed features is significantly higher than the correlation of the number of cycles.

[0027] Furthermore, the step 3 is to construct a lithium-ion battery state monitoring model based on random forest, including: randomly extracting samples from the original data set by the bootstrap method, training multiple decision trees, each tree independently predicts, and finally determining the final category Y according to the voting results of all trees.

[0028] Furthermore, the step 4 of establishing the state monitoring data set includes: extracting the battery features and constructing the battery state label, and establishing the data set for training and classification models.

[0029] Furthermore, the step 5 uses multiple evaluation indicators to evaluate the state monitoring model. Including: using different evaluation indicators to evaluate the results, accuracy (Accuracy), precision (Precision), recall rate (Recall), false alarm rate (FPR), false negative rate (FNR) and test time are used to evaluate the battery state monitoring results, using True Positive (TP), True Negative (TN), False Positive (FP) and False Negative (FN) are defined as follows:

[0030]

[0031] Furthermore, the step 6 builds an online status monitoring platform, including the following steps:

[0032] (1) Write the saved condition monitoring model in C language, use the relevant files including test data and model code to build the project and compile it in CCS6.0 platform.

[0033] (2) Burn the project into the DSP development board, obtain the time required for the status monitoring to run on the online platform, and display the relevant indicators on the development board.

[0034] Furthermore, the step 7 constructs a GRU-based lithium-ion battery health status assessment model. It includes: input feature F i The node status of the previous cycle is calculated through the control gate to obtain the current node status. After multiple calculations and status updates, the predicted value of the battery health status SOH' is finally output. k and by comparing the actual value SOH k Calculate loss, perform backpropagation and parameter update.

[0035] Furthermore, the step 8 of establishing a health status assessment data set includes: extracting the battery features and constructing the battery health status, and establishing the data set for training and predicting models.

[0036] Furthermore, the step 9 uses multiple evaluation indicators to evaluate the condition monitoring model, including: using different evaluation indicators such as root mean square error (RMSE), mean absolute error (MAE) and R square (R 2 ) To evaluate the prediction results, the evaluation indicators are defined as follows:

[0037]

[0038] Another object of the present invention is to provide a lithium-ion battery online state monitoring and health state assessment method and a lithium-ion battery online state monitoring and health state assessment system, comprising:

[0039] The battery status monitoring and health status assessment feature construction module is used to process and analyze the signals collected by the sensor during the charging and discharging process of the lithium-ion battery, extract the features based on the battery charging and discharging process, and use the feature data to complete the battery status monitoring and SOH assessment.

[0040] A battery status monitoring construction module is used to input feature data and corresponding status labels into a random forest model, train the random forest model, and build a battery status monitoring classification model;

[0041] The battery status monitoring classification module inputs the tested battery data into the trained random forest model to realize the status monitoring of lithium-ion batteries;

[0042] The battery online status monitoring construction module writes the saved battery status monitoring model in C language, uses the relevant files to build and compile the project in the CCS6.0 platform, burns the project to the DSP development board, and builds the battery online status monitoring platform;

[0043] A battery health status assessment building module is used to input feature data and corresponding battery health status into the GRU neural network, train the GRU neural network model, and build a battery health status assessment model;

[0044] The battery health status assessment prediction module inputs the tested battery data into the trained GRU neural network model to realize the health status assessment of the lithium-ion battery.

[0045] Another object of the present invention is to provide a computer device, the computer device includes a memory and a processor, 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 online status monitoring and health status assessment method.

[0046] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the lithium-ion battery online status monitoring and health status assessment method.

[0047] Another object of the present invention is to provide an information data processing terminal, which includes the lithium-ion battery online status monitoring and health status assessment system.

[0048] 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:

[0049] First, the present invention extracts relevant features based on the terminal voltage, current and temperature data of the complete cycle process during the battery charging and discharging process, and constructs battery status monitoring and health status assessment data sets respectively. The battery status monitoring data set is used and input into the random forest model to complete battery status monitoring. At the same time, considering the requirements for system real-time performance, online monitoring of the battery status is realized. The battery health status assessment data set is used and input into the GRU neural network to complete the battery health status assessment. The features used in the present invention come from the complete cycle process of the lithium-ion battery charging and discharging process, which can obtain more battery status change information and improve the accuracy of battery status monitoring and health status assessment.

[0050] The present invention takes into account the requirements of the battery management system for real-time status monitoring. Most of the current status monitoring solutions only achieve accurate battery status monitoring, but ignore the system's requirements for real-time status monitoring. In actual applications, the real-time performance of battery status monitoring is often crucial to the operating efficiency and safety of the system, and delayed response will affect the battery's fault warning and health management. The present invention takes into account the real-time requirements of lithium-ion battery status monitoring, builds an online status monitoring platform, and proves that the present invention can achieve real-time lithium-ion battery status monitoring and accurate health status assessment. After relevant adjustments, the present invention can be widely used in status monitoring and health status assessment of various types of lithium-ion batteries.

[0051] The lithium-ion battery state monitoring and health state assessment based on charging and discharging characteristics proposed in the present invention use random forest algorithm and GRU neural network respectively. The random forest algorithm is a set classifier containing multiple decision trees, which can effectively avoid the overfitting problem of classification results, improve the accuracy of battery state monitoring, and significantly improve the real-time performance of battery state monitoring on the online platform. The GRU neural network uses a gating mechanism to selectively memorize and forget past information to achieve the advantages of modeling sequence data and improve the accuracy of battery health state assessment. This design method is simple and easy to use, and can be widely used in different types of lithium-ion battery application scenarios. Compared with traditional lithium-ion battery health state assessment methods, this design method can monitor the battery state in real time and accurately, and complete high-precision health state assessment for batteries in normal state, so as to effectively reflect the battery usage for users. Therefore, this design method has strong innovation, applicability and a wide range of application fields.

[0052] Second, as auxiliary evidence of the inventiveness of the claims of the present invention, it is also reflected in the following important aspects:

[0053] (1) The expected benefits and commercial value of the technical solution of the present invention after transformation are:

[0054] The present invention realizes battery health system management by realizing online status monitoring and health status assessment of lithium-ion batteries, which can help users to accurately understand the battery status in real time, take corresponding measures in time, delay battery aging, reduce production problems caused by battery performance degradation, thereby reducing production costs and improving the operating reliability and stability of equipment. In the fields of consumer electronics, energy storage, aerospace, etc., the present invention manages and ensures the high efficiency and optimal performance of these systems, and ensures the stability and safety of lithium-ion battery operation.

[0055] (2) The technical solution of the present invention solves a technical problem that people have been eager to solve but have never been able to solve successfully:

[0056] The technical solution of the present invention considers the relevant data of the complete cycle of battery charging and discharging in feature extraction, and combines the online status monitoring model based on random forest and the health status assessment model based on GRU to improve the real-time performance and accuracy of the system. This comprehensive method helps users to fully and accurately understand the battery usage and realize battery health system management. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of a method for online status monitoring and health status assessment of a lithium-ion battery based on charge and discharge characteristics provided by an embodiment of the present invention.

[0058] Figure 2 It is a flow chart of an implementation method of an online state monitoring and health state assessment method for a lithium-ion battery based on charge and discharge characteristics provided in an embodiment of the present invention.

[0059] Figure 3 It is a schematic diagram of charge and discharge characteristics provided by an embodiment of the present invention.

[0060] Figure 4 It is a schematic diagram of the results of battery status monitoring using the random forest model provided in an embodiment of the present invention.

[0061] Figure 5 It is a schematic diagram of an online status monitoring system provided by an embodiment of the present invention.

[0062] Figure 6 It is a schematic diagram of the results obtained by performing battery health status assessment using the GRU neural network provided in an embodiment of the present invention.

[0063] Figure 7 It is a structural diagram of a lithium-ion battery online status monitoring and health status assessment system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0064] 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 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.

[0065] like Figure 1 As shown, the lithium-ion battery state monitoring and health state assessment method based on charge and discharge characteristics provided by the embodiment of the present invention includes the following steps:

[0066] S1, collects data collected by sensors during the charging and discharging process of the lithium-ion battery and pre-processes the data;

[0067] S2, extracting battery features according to the data preprocessed in step S1, constructing a battery state of health SOH, and performing correlation analysis between the battery features and the state label and the battery state of health SOH respectively;

[0068] S3, builds a lithium-ion battery status monitoring model based on the random forest model;

[0069] S4, establishing a state monitoring data set according to the preprocessed data, battery characteristics and battery state labels in step S1, and dividing the state monitoring data set into a training set and a test set, training the lithium-ion battery state monitoring model according to the training set, and testing the lithium-ion battery state monitoring model according to the test set;

[0070] S5, evaluating the classification result of the lithium-ion battery state monitoring model according to multiple evaluation indicators to verify the accuracy of the lithium-ion battery state monitoring model;

[0071] S6, using a digital signal processor DSP to construct an online state monitoring platform for lithium-ion batteries, and measuring the test time of the online state monitoring platform for lithium-ion batteries to verify the real-time performance of the online state monitoring platform;

[0072] S7, constructing a lithium-ion battery health status assessment model based on a gated recurrent unit neural network model;

[0073] S8, establishing a health status assessment data set according to the data preprocessed in step S1, the battery characteristics and the battery health status SOH, and dividing the health status assessment data set into a training set and a test set, training a lithium-ion battery health status assessment model according to the training set, and testing the lithium-ion battery health status assessment model according to the test set;

[0074] S9, evaluating the prediction result of the lithium-ion battery health status assessment model according to multiple evaluation indicators to verify the accuracy of the lithium-ion battery health status assessment model.

[0075] In a preferred embodiment of the present invention, the lithium-ion battery is repaired or replaced in a timely manner according to the real-time health status of the battery.

[0076] In a preferred embodiment of the present invention, step S2 extracts features based on the charging and discharging characteristic curves of the lithium-ion battery and performs mutual information and grey correlation analysis on the state label and the health state, respectively, and is specifically implemented as follows:

[0077] (1) Obtain data such as terminal voltage, current, and temperature during the charging and discharging process of lithium-ion batteries, and obtain the time required for a complete charge or discharge of the battery, the time required for the battery constant current charging process, the time required for the battery charging temperature to reach the highest point, and the time required for the battery discharge temperature to reach the mean point.

[0078] (2) Based on the obtained data, calculations are completed to obtain the constant current charging time proportion feature F1, the charging temperature reaching the highest point time proportion feature F2, and the discharge temperature reaching the mean time proportion F3.

[0079] (3) Process the battery capacity data and calculate the SOH, add status and SOH labels, use the mutual information method to calculate the correlation between the features and the status labels, and use the grey correlation analysis method to calculate the correlation between the features and the battery health status.

[0080] In a preferred embodiment of the present invention, the method for constructing a state monitoring model based on random forest in step S3 is as follows: a random forest prediction model is established, including a random forest structure of input features and output targets. In this model, the input features represent the constructed features, and the output targets represent the binary classification results. The random forest prediction model is trained using training data, and the classification of the battery state is achieved through an integrated learning mechanism of multiple decision trees, and finally an accurate monitoring result is obtained.

[0081] In a preferred embodiment of the present invention, the online state monitoring construction method of step S6 is: write the saved state monitoring model in C language, use test data and model code files to build a project and compile it in the CCS6.0 platform. Burn the project to the DSP development board, obtain the time required for the state monitoring to run on the online platform, and display the relevant indicators on the development board to realize online monitoring of the battery state.

[0082] In a preferred embodiment of the present invention, the method for constructing a health status assessment model based on a GRU neural network in step S7 is: design a GRU model, which includes an input layer, a GRU layer, a fully connected layer, and an output layer. In this model, the units of the input layer are used to represent the constructed time proportion features, and the units of the output layer are used to output the health status assessment results. By training the GRU model using training data, the prediction results of SOH are finally obtained.

[0083] The technical effects of the present invention are described in detail below in conjunction with tests.

[0084] The data used in the embodiment of the present invention comes from the lithium-ion battery aging data released by NASA Ames Research Center. The data set contains the operation experiments of lithium-ion batteries during the charging and discharging process. In order to study the real-time use of lithium-ion batteries, sensors collect different types of data. The data set records different types of data during the charging and discharging process of lithium-ion batteries, as well as the charging and discharging process data of lithium-ion batteries under different experimental environments.

[0085] Different lithium-ion battery data were selected to be divided into training sets and test sets, and the steps described in the present invention were used to perform online state monitoring and health state assessment of lithium-ion batteries. The experimental results obtained are as follows: Figure 4 To obtain the results using the random forest model, Figure 5 The results obtained from online status monitoring are: Figure 6 This is the result obtained using the GRU neural network model.

[0086] According to the experimental results, it can be seen that the online status monitoring and health status assessment of lithium-ion batteries by the method proposed in the present invention are closer to the actual use loss of lithium-ion batteries than the results obtained by the currently used prediction methods, and the time required for battery status monitoring is significantly reduced. Therefore, the method proposed in the present invention can better reflect the health status of lithium-ion batteries.

[0087] like Figure 7 As shown, the lithium-ion battery online status monitoring and health status assessment system provided by the embodiment of the present invention includes:

[0088] Battery status monitoring and health status assessment feature building module: used to process and analyze the signals collected by sensors during the charging and discharging process of lithium-ion batteries, extract features based on the battery charging and discharging process, and use feature data to complete battery status monitoring and health status assessment.

[0089] Battery status monitoring construction module: used to input feature data and corresponding status labels into the random forest model, train the random forest model, and build a battery status monitoring classification model;

[0090] Battery status monitoring classification module: input the tested battery data into the trained random forest model to realize the status monitoring of lithium-ion batteries;

[0091] Battery online status monitoring construction module: Write the saved battery status monitoring model in C language, use the relevant files to build and compile the project in the CCS6.0 platform, burn the project to the DSP development board, and build the battery online status monitoring platform;

[0092] Battery health status assessment building module: used to input feature data and corresponding battery health status into the GRU neural network, train the GRU neural network model, and build a battery health status assessment model;

[0093] Battery health status assessment prediction module: The tested battery data is input into the trained GRU neural network model to realize the health status assessment of the lithium-ion battery.

[0094] The application embodiment of the present invention utilizes the lithium-ion battery aging data of B0005, B0006, B0007 and B0018 at room temperature released by NASA Ames Research Center. First, the aging data of the battery charge and discharge process is preprocessed, and then the features are obtained to construct a battery state monitoring and health state assessment data set. The state monitoring data is divided into a training set and a test set, and the state monitoring model is trained using the training set data, and the model is tested using the test set data. An online state monitoring platform is constructed to realize online state monitoring. Finally, the health state assessment data is divided into a training set and a test set to complete the training and testing of the model.

[0095] The application examples of the present invention achieve good results in feature extraction and construction, and the grey correlation degrees of the features and SOH are both above 0.95.

[0096] The application embodiment of the present invention achieves good results in online status monitoring of lithium-ion batteries. Figure 6 The health status monitoring results of B0006 battery are shown when B0005 is used as the training set, with an accuracy of 0.9398 and the testing time on the online platform reduced to 156 microseconds.

[0097] The application embodiment of the present invention achieves good results in evaluating the health status of lithium-ion batteries. Figure 6 The health status assessment results of B0006 battery in normal state when B0005 is used as the training set are shown. 2 Reached 0.92.

[0098] 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 part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, 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. Such code is provided on the carrier medium. The device and its 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, etc., 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, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0099] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for online status monitoring and health status assessment of lithium-ion batteries, characterized in that: include: Step 1, collecting data collected by the sensor during the charging and discharging process of the lithium-ion battery, and preprocessing the data; Step 2: extract the battery features based on the data preprocessed in step 1, construct the battery health state SOH, and perform correlation analysis on the battery features, the status label and the battery health state SOH respectively; Step 3, building a lithium-ion battery status monitoring model based on the random forest model; Step 4, establishing a state monitoring data set according to the preprocessed data, battery characteristics and battery state labels in step 1, and dividing the state monitoring data set into a training set and a test set, training a lithium-ion battery state monitoring model according to the training set, and testing the lithium-ion battery state monitoring model according to the test set; Step 5, evaluating the classification results of the lithium-ion battery state monitoring model according to multiple evaluation indicators to verify the accuracy of the lithium-ion battery state monitoring model; Step 6, using a digital signal processor DSP to build an online state monitoring platform for lithium-ion batteries, and measuring the test time of the platform to verify the real-time performance of the online state monitoring platform; Step 7, constructing a lithium-ion battery health status assessment model based on the GRU neural network model; Step 8, establishing a health status assessment data set according to the preprocessed data, battery characteristics and battery health status SOH in step 1, and dividing the health status assessment data set into a training set and a test set, training a lithium-ion battery health status assessment model according to the training set, and testing the lithium-ion battery health status assessment model according to the test set; Step 9: Evaluate the prediction results of the lithium-ion battery health status assessment model according to multiple evaluation indicators to verify the accuracy of the lithium-ion battery health status assessment model.

2. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: The data collected by the sensor in step 1 include: voltage, current, and temperature, and the preprocessing includes processing of missing values, abnormal values, and noise.

3. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: The step 2 of extracting the characteristics of the lithium-ion battery includes: S21: Obtaining voltage and temperature data during the charging and discharging process of the lithium-ion battery, and obtaining the duration of the constant current charging phase of the battery, the time when the temperature reaches the highest point in the charging phase, the time when the temperature reaches the mean value in the discharging phase, the duration of the charging phase, and the duration of the discharging phase; S22: Use the obtained data to construct features based on the charging and discharging process, including the proportion of constant current time during the charging process F1, the proportion of battery temperature reaching peak time during the charging process F2, and the proportion of battery temperature reaching average time during the discharging process F3; obtain the battery capacity data and complete the construction of the battery health status, which is defined as follows: Among them, C k It indicates the remaining available capacity of the battery after the kth cycle, C0 indicates the rated capacity of the battery, and SOH k Indicates the battery health after the kth cycle; S23: Constructing a state monitoring and health status assessment data set; completing state division according to the battery capacity data to obtain a state label, a battery health of more than 80% is a normal state, marked as 0, and a health of less than 80% is an abnormal state, marked as 1, the extracted features and state labels are used to construct a state monitoring data set, and the features extracted under the normal state and the corresponding battery health state are used to construct a health status assessment data set; S24: Use mutual information analysis and grey correlation analysis to perform correlation analysis to obtain the correlation between the feature and status labels and the battery capacity. The formula of mutual information analysis is as follows: Among them, X is the proposed features F1, F2 and F3, and Y is the battery status label; the grey correlation method is defined as follows: Among them, ξ i (k) is feature F i The correlation coefficient of the kth cycle, r i F i The correlation degree with SOH, i = 1, 2, 3; ρ is the resolution coefficient, ρ∈(0, 1), ρ = 0.5, x i (k) is the proposed F i , N represents the battery cycle.

4. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: In step 3, a lithium-ion battery state monitoring model is constructed based on a random forest model, including: randomly extracting samples from the original data set by a bootstrap method, training multiple decision trees, each tree independently predicts, and finally determining the final category Y based on the voting results of all trees.

5. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: The step 4 of establishing a status monitoring data set includes: extracting the battery features and constructing the battery status label, and establishing the data set for training and classification models.

6. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: The multiple evaluation indicators in step 5 include accuracy, precision, recall, false alarm rate, missed alarm rate and test time.

7. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: The step 6 of establishing an online status monitoring platform for lithium-ion batteries includes: S61: Write the saved condition monitoring model in C language, use the relevant files including test data and model code to build the project and compile it in CCS6.0 platform; S62: Burn the project into the DSP development board, obtain the time required for the status monitoring to run on the online platform, and display the relevant indicators on the development board.

8. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: In step 7, a lithium-ion battery health status assessment model is constructed based on the GRU neural network model, including: input feature F i The node status of the previous cycle is calculated through the control gate to obtain the current node status. After multiple calculations and status updates, the predicted value of the battery health status SOH' is finally output. k and by comparing the actual value SOH k Calculate loss, perform backpropagation and parameter update.

9. The method for online status monitoring and health status assessment of a lithium-ion battery according to claim 1, characterized in that: In step 8, a health status assessment data set is established, including: extracting the battery features and constructing the battery health status, and establishing the data set for training and predicting models; in step 9, multiple evaluation indicators include root mean square error, mean absolute error and R square.

10. A lithium-ion battery online status monitoring and health status assessment system, characterized in that: The system comprises: Battery status monitoring and health status assessment feature building module, which is used to process and analyze the signals collected by sensors during the charging and discharging process of lithium-ion batteries, extract features based on the battery charging and discharging process, and use feature data to complete battery status monitoring and health status assessment; A battery status monitoring construction module is used to input feature data and corresponding status labels into a random forest model, train the random forest model, and build a battery status monitoring classification model; The battery status monitoring classification module is used to input the tested battery data into the trained random forest model to realize the status monitoring of lithium-ion batteries; The battery online status monitoring construction module is used to write the saved battery status monitoring model in C language, use the relevant files to build and compile the project in the CCS6.0 platform, burn the project to the DSP development board, and build the battery online status monitoring platform; A battery health status assessment building module is used to input feature data and corresponding battery health status into the GRU neural network, train the GRU neural network model, and build a battery health status assessment model; The battery health status assessment prediction module is used to input the tested battery data into the trained GRU neural network model to realize the health status assessment of the lithium-ion battery.

Images