A battery health assessment method and system

Abstract

The present disclosure provides a battery health degree evaluation method and system, and relates to the technical field of battery management. The method comprises the following steps: obtaining target battery operation data; inputting the target battery operation data into a trained parameter prediction model to obtain first prediction data; inputting target battery physical and chemical characteristic parameters into a trained battery physical model to obtain target battery physical and chemical characteristics; inputting the first prediction data into the battery physical model to obtain updated target battery physical and chemical characteristics; inputting the updated target battery physical and chemical characteristics into the parameter prediction model to obtain second prediction data; and determining the health degree of the target battery according to the second prediction data. The present disclosure can improve the accuracy of battery health degree evaluation.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of battery management, and in particular to a battery health degree evaluation method and system. BACKGROUND

[0002] The storage capacity and the rapid charging and discharging capacity of the power battery will continuously decrease with aging. The most direct performance of the aging of the power battery is the decrease of the releasable energy of the power battery and the decrease of the power level, which is specifically the attenuation of the battery capacity and the increase of the internal resistance. The health degree (SOH) is a quantitative index for evaluating the health degree of the power battery. Generally, the SOH of the new power battery is initially set to 100%, and when the capacity of the power battery decreases to 80% of the initial capacity, it is considered that the power battery does not meet the normal demand.

[0003] In the related art, the future time parameter of the battery can be predicted based on the historical running data of the battery, and the health degree of the battery can be obtained according to the future time parameter. However, this method only uses the historical running data of the battery for prediction, and the accuracy needs to be improved. SUMMARY

[0004] The embodiments of the present disclosure provide a battery health degree evaluation method and system, which can improve the accuracy of battery health degree evaluation. The technical solutions are as follows:

[0005] In a first aspect, a battery health degree evaluation method is provided, comprising:

[0006] obtaining target battery running data;

[0007] inputting the target battery running data into a trained parameter prediction model to obtain first prediction data;

[0008] inputting the physical and chemical characteristic parameters of the target battery into a trained battery physical model to obtain the physical and chemical characteristics of the target battery; inputting the first prediction data into the battery physical model to obtain updated physical and chemical characteristics of the target battery;

[0009] inputting the updated physical and chemical characteristics of the target battery into the parameter prediction model to obtain second prediction data;

[0010] determining the health degree of the target battery according to the second prediction data.

[0011] In a second aspect, a battery health degree evaluation system is provided, comprising:

[0012] a data acquisition module configured to acquire target battery running data;

[0013] The first parameter prediction module is configured to input the target battery operation data into the trained parameter prediction model to obtain first prediction data.

[0014] The physical and chemical characteristic simulation module is configured to input the target battery physical and chemical characteristic parameter into the trained battery physical model to obtain target battery physical and chemical characteristics, and input the first prediction data into the battery physical model to obtain updated target battery physical and chemical characteristics.

[0015] The second parameter prediction module is configured to input the updated target battery physical and chemical characteristics into the parameter prediction model to obtain second prediction data.

[0016] The health degree evaluation module is configured to determine the health degree of the target battery according to the second prediction data.

[0017] In a third aspect, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, and the processor executes the computer program to complete the steps of the above battery health degree evaluation method.

[0018] In a fourth aspect, a computer readable storage medium is provided for storing computer instructions, and the computer instructions are executed by a processor to complete the steps of the above battery health degree evaluation method.

[0019] In a fifth aspect, a computer program product is provided, including computer programs / instructions, and the computer programs / instructions are executed by a processor to implement the steps of the above battery health degree evaluation method.

[0020] The technical scheme provided by the embodiments of the present disclosure has the beneficial effects that: in the embodiments of the present disclosure, by combining the parameter prediction model and the battery physical model, the accurate evaluation of the battery health degree is realized. The parameter prediction model is used to predict the battery data at the next time according to the historical data of the target battery; and the battery physical model is used to simulate the physical and chemical characteristics of the target battery. The first prediction data obtained by the parameter prediction model is input into the battery physical model, the physical and chemical characteristics of the target battery are updated, and then the updated physical and chemical characteristics are fed back to the parameter prediction model to update the weight and bias of the parameter prediction model, so as to obtain the second prediction data for determining the battery health degree. Through this feedback mechanism, the parameter prediction model can more accurately learn the aging characteristics of the battery, improve the prediction accuracy, and further improve the accuracy of the health degree evaluation.

[0021] The advantages of the additional aspects of the present disclosure will be partially given in the following description, partially will become obvious from the following description, or will be understood by the practice of the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings needed to be used in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0023] Figure 1 is a flowchart of a battery health degree evaluation method provided by an embodiment of the present disclosure;

[0024] Figure 2 is a structural block diagram of a battery health degree evaluation system provided by an embodiment of the present disclosure;

[0025] Figure 3 is a structural block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0026] In order to make the objects, technical solutions and advantages of the present disclosure clearer, the embodiments of the present disclosure will be further described in detail below with reference to the drawings.

[0027] Figure 1 is a flowchart of a battery health degree evaluation method provided by an embodiment of the present disclosure, referring to Figure 1 , the method comprises:

[0028] Step 101, obtaining target battery operation data;

[0029] Step 102, inputting the target battery operation data into a trained parameter prediction model to obtain first prediction data;

[0030] Step 103, inputting the target battery physical and chemical characteristic parameters into a trained battery physical model to obtain target battery physical and chemical characteristics; inputting the first prediction data into the battery physical model to obtain updated target battery physical and chemical characteristics;

[0031] Step 104, inputting the updated target battery physical and chemical characteristics into the parameter prediction model to obtain second prediction data;

[0032] Step 105, determining the health degree of the target battery according to the second prediction data.

[0033] In step 101, the collection and preprocessing of data are first performed. The battery data is collected in real time by the battery management software in the vehicle controller, specifically including voltage, current, temperature, SOC (state of charge), capacity, charge cycle number, battery cell serial number, etc. Subsequently, the collected data is subjected to abnormality detection and preprocessing, and the missing or obviously abnormal data is filtered, and then the data is subjected to Z-value centralization and normalization processing, thereby accelerating the speed of subsequent model training and optimization. The abnormality detection can be realized by using existing abnormality detection algorithms, such as 3σ method, Z-value method, k-means clustering algorithm, etc. The data is collected and updated to the database in real time for subsequent prediction.

[0034] In this embodiment, the database and the model are preferably deployed in the cloud to fully utilize the parallel processing capability of cloud computing, accelerate the model training and inference speed, and realize large-scale data processing. In some embodiments, to meet the real-time requirement and improve data security, the model can also be deployed on the vehicle side. After the model training is completed in the cloud or on the vehicle side, the model is imported into the vehicle side to realize the prediction task locally.

[0035] The edge computing node is deployed on the vehicle side to perform preliminary data processing and reduce the transmission burden. The preliminary data processing refers to determining the data that needs to be uploaded according to the prediction requirement. For example, the vehicle-side data is collected at a step length of 0.1 s, and the cloud-side prediction is performed at a step length of 1 s. Therefore, the vehicle side only needs to upload the data once every 1 s, and does not need to upload all the data. At this time, the vehicle side can perform data processing, such as merging the data or selecting a group of data for uploading.

[0036] The processed data is transmitted to the whole vehicle T-box through the CAN network, the T-box communicates with the cloud, and the data is transmitted to the database in the cloud. After the data in the database is subjected to feature extraction, it is imported into the cloud model for processing.

[0037] Before step 102 is performed, the parameter prediction model and the battery physical model are also trained and deployed. The model deployed in the cloud is a health degree prediction model combined with the parameter prediction model and the battery physical model. The parameter prediction model selects the LSTM (Long Short-Term Memory) model, and the battery physical model selects the P2D (Pseudo-2-Dimensional Model) model. The LSTM model is used to process time series data, and the P2D model is used to simulate the physical and chemical behavior of the battery.

[0038] The LSTM model is used to predict the next step data according to the input historical data, which is a multi-dimensional single-step network architecture. The input of the LSTM can be represented as [samples, timesteps, features], where samples represent the amount of data, timesteps represent the time step, i.e., the prediction of the next step data using the historical data of the previous steps, and features represent the feature dimension of the data.

[0039] The P2D model considers the physical and chemical processes inside the battery, including electrode reactions, ion transport, electrolyte concentration gradient, etc. The P2D model can simulate the voltage response and capacity change of the battery, and provide mechanism analysis of battery aging. Model structure: The P2D model includes electrode equations, ion transport equations, electrolyte equations, etc., which describe the multi-physical field coupling process inside the battery. Electrode equation: describes the lithium ion intercalation and deintercalation reaction in the electrode material, and the electrode reaction kinetics is characterized by the Butler-Volmer equation. Ion transport equation: describes the diffusion and migration of lithium ions in the electrolyte, considering the influence of concentration gradient and electric field. Electrolyte equation: describes the ion concentration distribution and its change in the electrolyte, and the ion flow is described by the Nernst-Planck equation. Boundary conditions and initial conditions: set the boundary conditions and initial conditions of the battery under different working conditions, simulate the actual use scenario.

[0040] The model training process includes:

[0041] Step 1011, initialize the parameters of the LSTM model and the P2D model. The input of the LSTM model is the historical running data of the battery (voltage, current, temperature, SOC, capacity, charge cycle number, battery monomer serial number, etc.), and the input of the P2D model is the physical and chemical characteristic parameters of the battery, including battery material parameters (positive and negative material parameters, electrolyte parameters), geometric parameters (electrode thickness, separator thickness, volume ratio of electrode and separator), kinetic parameters (electrode reaction rate constant, transfer coefficient), operating conditions (charge and discharge current, temperature, SOC), other parameters (physical properties of electrode and electrolyte).

[0042] Step 1012, train the LSTM model using the training set data, extract time series features, and update the model weights. The LSTM model can learn the behavior characteristics of the battery under different working conditions and provide time series prediction results. The battery historical data is divided into training set and test set, where the training set accounts for 90% of the total data and the test set accounts for 10%. The training set is used to train the model parameters, and the test set is used to measure the goodness of the model. After dividing the data set, the data is preprocessed (here, Z value centralization processing is adopted), which can speed up the gradient descent to find the optimal solution.

[0043] Step 1013, simulate the physical and chemical characteristics of the battery using the P2D model, obtain the physical and chemical characteristics of the battery, including the voltage-time curve of charging and discharging, the internal state (electrolyte concentration, solid-phase lithium ion concentration) distribution, current density distribution, temperature distribution, charging and discharging capacity, battery impedance, OCV, SOC, polarization voltage, reaction overpotential, battery efficiency. The P2D model can simulate the physical and chemical processes inside the battery in detail and provide the response characteristics of the battery under different aging states.

[0044] Step 1014, use the prediction results of the LSTM model as the input of the P2D model for secondary correction. Specifically, input the capacity data predicted by the LSTM model at the next time into the P2D model, and the P2D model further simulates the physical and chemical behavior of the battery to correct the error of the LSTM prediction.

[0045] Step 1015, feed back the SOC and capacity data output by the P2D model to the LSTM model, compare with the prediction results of the LSTM model, and update the weights and biases of the LSTM model using the SOC and capacity data output by the P2D model. Through this feedback mechanism, the LSTM model can more accurately learn the aging characteristics of the battery and improve the prediction accuracy.

[0046] Step 1016, use the Adam optimizer to optimize the combined model (i.e. the LSTM model and the P2D model) to reduce the prediction error. The parameters of the LSTM model and the P2D model are optimized together to ensure the accuracy and stability of the model.

[0047] Step 1017, use test set data to evaluate the performance of the model, adjust the model parameters, and ensure the generalization ability of the model. Through the validation of the test data, the prediction accuracy and robustness of the combined model are evaluated.

[0048] In step 102, input the target battery operation data in the database into the parameter prediction model (the LSTM model is used in this embodiment), and use the trained LSTM model to predict the first prediction data, which is the predicted voltage, current, temperature, SOC, capacity, charging cycle number, etc. data at the next time.

[0049] In step 103, input the physical and chemical characteristics of the target battery in the database into the battery physical model (the P2D model is used in this embodiment), and simulate the physical and chemical characteristics of the target battery. Then input the first prediction data predicted in step 102 into the battery physical model to update the simulation results and obtain the updated physical and chemical characteristics of the target battery.

[0050] In step 104, the SOC and capacity data in the updated target battery physical and chemical characteristics obtained by the battery physical model are fed back into the parameter prediction model to replace the prediction results of the parameter prediction model for the SOC and capacity data, the weights and biases of the parameter prediction model are updated, and then the prediction results output by the updated model are used as the second prediction data.

[0051] In step 105, the health degree of the target battery can be obtained by comparing the capacity data in the second prediction data with the rated capacity of the target battery. The calculation formula is as follows:

[0052]

[0053] Wherein, SOH represents the health degree, Q n represents the current capacity, Q d represents the rated capacity.

[0054] The method provided above realizes an adaptive learning mechanism, can dynamically adjust model parameters according to newly collected data, improves the real-time adaptability of the model, and the calculation ability and accuracy of the model will gradually increase with the accumulation of historical data. Based on the collaborative prediction of the LSTM model and the P2D model, compared with simply using the prediction model for prediction, the physical and chemical characteristics of the battery are introduced to feedback and update the model, thereby improving the prediction accuracy. Based on the prediction result of SOH, predictive maintenance suggestions can be further provided to prevent battery failure, prolong the service life of the battery, and further introduce the links of battery recycling and reuse, and determine the secondary utilization value of the recycled battery through SOH evaluation.

[0055] In addition, with the progress of the vehicle controller capability and the strengthening of the communication capability, the vehicle cloud combined technology in the embodiment has the possibility of engineering practice. The model and the database in the embodiment are both arranged in the cloud, and the high calculation requirement of the data volume can be overcome by the cloud hardware, so that the parallel processing capability of the cloud computing can be utilized to speed up the model training and inference speed and support large-scale data processing. Therefore, safety and reliability also need to be considered. For data security, data encryption and access control mechanisms are implemented to ensure the safety of data transmission and storage between the vehicle and the cloud; for system reliability, a redundant system and a backup mechanism are constructed to ensure that the system can still operate stably in unexpected situations.

[0056] Figure 2 is a structural block diagram of a structure 200 of a battery health degree evaluation system provided by the embodiment of the disclosure, as Figure 2 shown, the system comprises a data acquisition module 201, a first parameter prediction module 202, a physical and chemical characteristic simulation module 203, a second parameter prediction module 204, and a health degree evaluation module 205.

[0057] The data acquisition module 201 is configured to acquire target battery operation data.

[0058] The first parameter prediction module 202 is configured to input the target battery operation data into a trained parameter prediction model to obtain first prediction data.

[0059] The physical and chemical characteristic simulation module 203 is configured to input target battery physical and chemical characteristic parameters into a trained battery physical model to obtain target battery physical and chemical characteristics, and input the first prediction data into the battery physical model to obtain updated target battery physical and chemical characteristics.

[0060] The second parameter prediction module 204 is configured to input the updated target battery physical and chemical characteristics into the parameter prediction model to obtain second prediction data.

[0061] The health degree evaluation module 205 is configured to determine the health degree of the target battery according to the second prediction data.

[0062] It should be noted that the battery health degree evaluation system 200 provided in the above embodiment is only used as an example for the division of the above functional modules during health degree evaluation. In actual applications, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the battery health degree evaluation system 200 and the battery health degree evaluation method provided in the above embodiment belong to the same concept, and the specific implementation process is described in the method embodiment, which will not be repeated here.

[0063] Figure 3 is a structural block diagram of an electronic device provided by the embodiments of the present disclosure. As shown in Figure 3 the electronic device 300 can be a vehicle-mounted computer or the like. The electronic device 300 includes a processor 301 and a memory 302.

[0064] The processor 301 can include one or more processing cores, such as a 4-core processor, an 8-core processor, and the like. The processor 301 can be implemented in at least one of a hardware form of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), an LA (Programmable Logic Array). The processor 301 can also include a main processor and a coprocessor, the main processor being a processor for processing data in an awake state, also referred to as a CPU (Central Processing Unit), and the coprocessor being a low-power processor for processing data in a standby state. In some embodiments, the processor 301 can be integrated with a GPU (Graphics Processing Unit) for rendering and drawing content required to be displayed by a display screen. In some embodiments, the processor 301 can further include an AI (Artificial Intelligence) processor for processing computing operations related to machine learning.

[0065] The memory 302 can include one or more computer-readable media, which can be non-transitory. The memory 302 can also include a high-speed random access memory, and a nonvolatile memory such as one or more disk storage devices, flash storage devices. In some embodiments, the non-transitory computer-readable medium in the memory 302 is used to store at least one computer program for being executed by the processor 301 to implement the battery health evaluation method provided in the embodiments of the present disclosure.

[0066] Those skilled in the art can understand that the structure shown in the figure does not constitute a limitation on the electronic device 300, and can include more or fewer components than those shown, or combine certain components, or adopt different component arrangements. Figure 3 The structure shown in the figure does not constitute a limitation on the electronic device 300, and can include more or fewer components than those shown, or combine certain components, or adopt different component arrangements.

[0067] The embodiments of the present disclosure further provide a computer-readable storage medium for storing computer instructions, which can complete the steps of the battery health evaluation method provided in the embodiments of the present disclosure when the computer instructions are executed by a processor.

[0068] The embodiments of the present disclosure further provide a computer program product, including computer programs / instructions, which implement the steps of the battery health evaluation method provided in the embodiments of the present disclosure when executed by a processor.

[0069] The above merely provides preferred embodiments of the present disclosure, and is not used to limit the present disclosure. For those skilled in the art, the present disclosure can have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the principles and technical scope of the present disclosure shall fall into the scope of the present disclosure.

Claims

1. A method of evaluating a state of health of a battery, characterized by, The method comprises: acquiring target battery operation data; inputting the target battery operation data into a trained parameter prediction model to obtain first prediction data; inputting target battery physical and chemical characteristic parameters into a trained battery physical model to obtain target battery physical and chemical characteristics; inputting the first prediction data into the battery physical model to obtain updated target battery physical and chemical characteristics; inputting the updated target battery physical and chemical characteristics into the parameter prediction model to obtain second prediction data, specifically: inputting SOC and capacity data into the parameter prediction model to update the weight and bias of the parameter prediction model, and outputting the second prediction data using the updated parameter prediction model; determining the health degree of the target battery according to the second prediction data; the parameter prediction model is a long short-term memory network model, and the battery physical model is a pseudo two-dimensional model.

2. The method of claim 1, wherein, The target battery operation data comprises voltage, current, temperature, state of charge, capacity, charge cycle number, and battery cell serial number of the target battery.

3. The method of claim 1, wherein, The first prediction data comprises voltage, current, temperature, SOC, capacity, and charge cycle number of the target battery at the next time point.

4. The method of claim 1, wherein, The capacity data in the second prediction data is divided by the rated capacity of the target battery to obtain the health degree of the target battery.

5. A battery state of health assessment system, comprising: The method comprises: a data acquisition module configured to acquire target battery operation data; a first parameter prediction module configured to input the target battery operation data into a trained parameter prediction model to obtain first prediction data; a physical and chemical characteristic simulation module configured to input target battery physical and chemical characteristic parameters into a trained battery physical model to obtain target battery physical and chemical characteristics; and input the first prediction data into the battery physical model to obtain updated target battery physical and chemical characteristics; a second parameter prediction module configured to input the updated target battery physical and chemical characteristics into the parameter prediction model to obtain second prediction data, specifically: inputting SOC and capacity data into the parameter prediction model to update the weight and bias of the parameter prediction model, and outputting the second prediction data using the updated parameter prediction model; a health degree evaluation module configured to determine the health degree of the target battery according to the second prediction data; the parameter prediction model is a long short-term memory network model, and the battery physical model is a pseudo two-dimensional model.

6. An electronic device, comprising: The computer program is executed by the processor to complete the steps of the method of any one of claims 1-4.

7. A computer readable storage medium characterized in that, The computer program is executed by the processor to complete the steps of the method of any one of claims 1-4.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to complete the steps of the method of any one of claims 1-4.

Images