Lithium ion battery health state estimation method fusing physical information

Abstract

The invention relates to a lithium ion battery health state estimation method fusing physical information, and belongs to the technical field of batteries, and the method comprises the following steps: S1, collecting the operation data of new energy vehicles of the same vehicle type, including charging and discharging data, and building a new energy vehicle operation database; s2, performing health feature extraction based on the battery charging data and calculating a capacity label; s3, establishing a lithium ion battery health state estimation model based on the feedforward neural network, defining a loss function fusing physical information, and training the model by using a gradient descent method; and S4, estimating the health state of the lithium ion battery of the new energy automobile based on the trained model. According to the method, the defects of poor interpretability and weak feature utilization capability of a deep learning black box method are overcome; and meanwhile, the advantage that a neural network model effectively extracts high-dimensional abstract features and the advantage that the model can correctly understand the potential influence of physical information features on battery aging are combined.

Description

Technical Field

[0001] The invention belongs to the technical field of batteries and relates to a method for estimating the health status of a lithium-ion battery integrating physical information. Background Art

[0002] The state of health (SOH) estimation methods for lithium-ion batteries of new energy vehicles can generally be divided into three types: ampere-hour integration method, model-based method and data-driven method. The ampere-hour integration method consumes a lot of time and energy and cannot be widely used in the SOH estimation scenario of new energy vehicle batteries; the model-based method relies on high-precision battery models and reliable signal sampling, which is difficult to apply in complex scenarios; the data-driven method extracts health features related to battery aging and constructs a mapping model between health features and SOH to achieve SOH estimation. This method has wide applicability, high estimation accuracy and high computational efficiency, and has received widespread attention in recent years. However, the conventional machine learning model that the data-driven method relies on is a black box model with poor model interpretability and weak feature utilization capabilities. It is often unable to correctly understand and utilize health features containing high-value physical information, which reduces the accuracy of the model's SOH estimation in complex scenarios. At present, no effective battery SOH estimation method that integrates physical information has been proposed to accurately estimate the SOH of lithium-ion batteries of new energy vehicles. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a method for estimating the health status of lithium-ion batteries of new energy vehicles by integrating physical information, which can achieve accurate estimation of the health status of lithium-ion batteries of new energy vehicles considering physical information.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] A method for estimating the health status of a lithium-ion battery by integrating physical information comprises the following steps:

[0006] S1: Collect the operation data of new energy vehicles of the same model, including charging and discharging data, and establish a new energy vehicle operation database;

[0007] S2: Extract health features based on battery charging data and calculate capacity labels;

[0008] S3: Establish a lithium-ion battery health status estimation model based on a feedforward neural network, define a loss function that integrates physical information, and use the gradient descent method to train the model;

[0009] S4: Estimate the health status of lithium-ion batteries for new energy vehicles based on the trained model.

[0010] Further, the step S1 specifically includes the following steps:

[0011] S11: Collect the operating data of a certain electric vehicle, including charging and discharging current, battery cell voltage, battery cell temperature, time, mileage and state of charge SOC;

[0012] S12: Establish an operation database of a certain car based on the collected vehicle operation data.

[0013] Further, the step S2 is specifically as follows:

[0014] S21: analyzing charging data of the electric vehicle, and selecting charging segments that satisfy the conditions that the charging start SOC is less than a first percentage and the charging end SOC is a second percentage as capacity calculation segments;

[0015] S22: Calculate the charging capacity of the capacity calculation segment by the ampere-hour integration method, and divide it by the SOC change of the capacity calculation segment to obtain a capacity label;

[0016] S23: Divide the capacity label by the rated capacity of the battery to obtain a health status SOH label;

[0017] S24: calculating an average cell voltage sequence, a highest cell voltage sequence, and a lowest cell voltage sequence in the capacity calculation segment according to the battery cell voltages;

[0018] S25: selecting a characteristic start voltage and a characteristic cut-off voltage, and extracting a characteristic extraction segment from the capacity calculation segment according to the average cell voltage sequence, wherein the starting point of the characteristic extraction segment is the moment when the average cell voltage sequence reaches the characteristic start voltage, and the ending point of the characteristic extraction segment is the moment when the average cell voltage sequence reaches the characteristic cut-off voltage;

[0019] S26: Calculate the cumulative charged power of the feature extraction segment by the ampere-hour integration method as health feature 1, calculate the standard deviation of the power increment sequence of the feature extraction segment by the ampere-hour integration method as health feature 2, calculate the highest single cell voltage at the starting point of the feature extraction segment as feature 3, calculate the lowest single cell voltage at the starting point of the feature extraction segment as feature 4, calculate the highest single cell voltage at the end point of the feature extraction segment as feature 5, calculate the lowest single cell voltage at the end point of the feature extraction segment as feature 6, calculate the range of single cell voltages at the starting point of the feature extraction segment as feature 7, calculate the range of single cell voltages at the end point of the feature extraction segment as feature 8, calculate the average single cell temperature of the feature extraction segment as feature 9, and extract the cumulative mileage of the vehicle as feature 10.

[0020] Further, the step S3 is specifically as follows:

[0021] S31: A lithium-ion battery SOH estimation model is established based on a multi-layer fully connected neural network. The input of the model is features 1 to 10, and the output of the model is the SOH label.

[0022] S32: Calculate the partial differential dT of the model output value with respect to feature 9, and define the physical loss 1L1 of the model as the larger of 0 and -dT;

[0023] S33: Calculate the partial differential dM of the model output value with respect to the feature 10, and define the physical loss 2L2 of the model as the larger of 0 and dM;

[0024] S34: Calculate the mean absolute percentage error (MAPE) of the model output value and the SOH label, and define MAPE as the estimated loss L3 of the model;

[0025] S35: Define the total loss L of the model as the weighted sum of L1, L2, and L3;

[0026] S36: Iteratively train the neural network model according to L and the gradient descent algorithm until L drops to a preset value or there is no longer a downward trend.

[0027] Further, the step S4 is specifically as follows:

[0028] Step S41: extracting features 1 to 10 from the charging data of the vehicle to be detected;

[0029] Step S42: Input the features 1 to 10 of the vehicle to be detected into the model, and the model outputs an estimated value of SOH.

[0030] The beneficial effects of the present invention are:

[0031] 1) Extracted health features that can effectively reflect the battery aging status from the charging data of new energy vehicle lithium-ion batteries;

[0032] 2) A neural network model was constructed, and physical information was integrated into the model from the perspective of model training, which solved the shortcomings of deep learning black box methods, such as poor interpretability and weak feature utilization capabilities;

[0033] 3) The proposed scheme combines the advantages of the neural network model in effectively extracting high-dimensional abstract features and the model's ability to correctly understand the potential impact of physical information characteristics on battery aging.

[0034] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0036] Figure 1 A flow chart of the overall method of the present invention;

[0037] Figure 2 It is the overall framework diagram of the embodiment method. DETAILED DESCRIPTION

[0038] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0039] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0040] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0041] A health status estimation method for lithium-ion batteries of new energy vehicles integrating physical information.

[0042] See also Figure 1 ,A method for estimating the health state of lithium ion batteries for new energy vehicles integrating physical information can be divided into the following steps:

[0043] Step S1: Collecting the operation data of new energy vehicles of the same model, including charging and discharging data, and establishing a new energy vehicle operation database;

[0044] Step S2: extracting health features based on battery charging data and calculating capacity labels;

[0045] Step S3: establishing a lithium-ion battery health state estimation model based on a feedforward neural network, defining a loss function integrating physical information, and training the model using a gradient descent method;

[0046] Step S4: estimating the health status of the lithium-ion battery of the new energy vehicle based on the trained model;

[0047] As an optional embodiment, the complete technical roadmap of this solution is as follows: Figure 2 shown.

[0048] As an optional embodiment, the above step S1 specifically includes steps S11-S12:

[0049] Collect the operating data of a certain electric vehicle, including charging and discharging current, battery cell voltage, battery cell temperature, time, mileage and state of charge (SOC);

[0050] Step S12: Establishing an operation database of a certain car based on the collected vehicle operation data.

[0051] As an optional embodiment, the above step S2 specifically includes steps S21-S26:

[0052] Step S21: analyzing the charging data of the electric vehicle, and selecting charging segments that meet the conditions that the charging start SOC is less than 20% and the charging end SOC is 100% as capacity calculation segments;

[0053] Step S22: Calculate the charging capacity of the capacity calculation segment by the ampere-hour integration method, and divide it by the SOC change of the capacity calculation segment to obtain the capacity label Ca, as shown in the following formula:

[0054]

[0055] Where I is the charging current, SOC e End SOC for capacity calculation segment, SOC s The starting SOC of the capacity calculation segment; Step S23: Divide the capacity tag by the rated capacity of the battery to obtain the health state SOH, as shown in the following formula:

[0056]

[0057] Where, Ca rated is the rated capacity of the battery;

[0058] Step S24: calculating the average cell voltage sequence, the highest cell voltage sequence, and the lowest cell voltage sequence of the capacity calculation segment according to the battery cell voltages;

[0059] Step S25: Select a characteristic start voltage and a characteristic cutoff voltage, and extract a characteristic extraction segment from the capacity calculation segment according to the average monomer voltage sequence, wherein the starting point of the characteristic extraction segment is the moment when the average monomer voltage sequence reaches the characteristic start voltage, and the end point of the characteristic extraction segment is the moment when the average monomer voltage sequence reaches the characteristic cutoff voltage.

[0060] As an optional embodiment, based on charging data statistics and user behavior analysis, the characteristic starting voltage and the characteristic cutting-off voltage are selected, specifically including:

[0061] The average cell voltage at the beginning of charging of a large number of charging segments of the same model was calculated to be 3380mV. The average cell voltage at the beginning of charging of more than 80% of the charging behaviors was lower than 3400mV.

[0062] The average cell voltage at the end of charging for a large number of charging segments of the same model was calculated to be 3600mV. The average cell voltage at the beginning of more than 80% of the charging behaviors was higher than 3580mV.

[0063] The feature start voltage is selected as 3400mV and the feature cutoff voltage is selected as 3580mV. Most users' charging behaviors span this range, which meets the conditions for extracting healthy features. Extracting healthy features from this range ensures the versatility of the method.

[0064] Step S26: Calculate the cumulative charged power of the feature extraction segment by the ampere-hour integration method as health feature 1, calculate the standard deviation of the power increment sequence of the feature extraction segment by the ampere-hour integration method as health feature 2, calculate the highest single cell voltage at the starting point of the feature extraction segment as feature 3, calculate the lowest single cell voltage at the starting point of the feature extraction segment as feature 4, calculate the highest single cell voltage at the end point of the feature extraction segment as feature 5, calculate the lowest single cell voltage at the end point of the feature extraction segment as feature 6, calculate the cell voltage range at the starting point of the feature extraction segment as feature 7, calculate the cell voltage range at the end point of the feature extraction segment as feature 8, calculate the average cell temperature of the feature extraction segment as feature 9, and extract the cumulative mileage of the vehicle as feature 10.

[0065] As an optional embodiment, the above step S3 specifically includes steps S31-S36:

[0066] Step S31: establishing a lithium-ion battery SOH estimation model based on a multi-layer fully connected neural network, the input of the model is features 1 to 10, and the output of the model is a SOH label;

[0067] Step S32: Calculate the partial differential (dT) of the model output value with respect to feature 9, and define the physical loss 1 (L1) of the model as the larger of 0 and -dT. The calculation method is as follows:

[0068]

[0069] L1=max(0,-dT)

[0070] In the formula, is the estimated SOH value output by the model, T is feature 9 (the average monomer temperature of the feature extraction segment), and this physical loss causes the estimated SOH value to increase with increasing temperature;

[0071] Step S33: Calculate the partial differential (dM) of the model output value with respect to the feature 10, and define the physical loss 2 (L2) of the model as the larger of 0 and dM. The calculation method is as follows:

[0072]

[0073] L2=max(0,dM)

[0074] Where M is feature 10 (the accumulated mileage of the vehicle). This physical loss causes the estimated SOH value to decrease as the vehicle mileage increases.

[0075] Step S34: Calculate the mean absolute percentage error (MAPE) between the model output value and the SOH label, and define MAPE as the estimated loss (L3) of the model. Other error calculation methods may also be used, such as the root mean square error (RMSE);

[0076] Step S35: define the total loss L of the model as the weighted sum of L1, L2, and L3, as shown in the following formula:

[0077] L=α*L1+β*L2+γ*L3

[0078] In the formula, α, β, and γ are weight terms, which can be adjusted according to actual needs;

[0079] Step S36: Iteratively train the neural network model according to L and the gradient descent algorithm until L drops to a preset value or no longer has a downward trend.

[0080] As an optional embodiment, the above step S4 specifically includes steps S41-S42:

[0081] Step S41: extracting features 1 to 10 from the charging data of the vehicle to be detected;

[0082] Step S42: Input the features 1 to 10 of the vehicle to be detected into the model, and the model outputs an estimated value of SOH.

[0083] In the above embodiments, the description's reference to "this embodiment" indicates that a particular feature, structure, or characteristic described in conjunction with the embodiment is included in at least some embodiments, but not necessarily all embodiments. Multiple occurrences of "this embodiment" do not necessarily all refer to the same embodiment.

[0084] In the above-described embodiments, although the invention has been described in conjunction with specific embodiments of the invention, many substitutions, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. For example, other storage structures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed. Embodiments of the invention are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of the appended claims.

[0085] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the methods in this embodiment is implemented.

[0086] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0087] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0088] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0089] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0090] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0091] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0092] The present invention can be used in many general or special computing system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0093] The present invention may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A method for estimating the health status of a lithium-ion battery by integrating physical information, characterized in that: The following steps are involved: S1: Collect the operation data of new energy vehicles of the same model, including charging and discharging data, and establish a new energy vehicle operation database; S2: Extract health features based on battery charging data and calculate capacity labels; S3: Establish a lithium-ion battery health status estimation model based on a feedforward neural network, define a loss function that integrates physical information, and use the gradient descent method to train the model; S4: Estimate the health status of lithium-ion batteries for new energy vehicles based on the trained model.

2. The method for estimating the health status of a lithium-ion battery integrating physical information according to claim 1, characterized in that: The step S1 specifically includes the following steps: S11: Collect the operating data of a certain electric vehicle, including charging and discharging current, battery cell voltage, battery cell temperature, time, mileage and state of charge SOC; S12: Establish an operation database of a certain car based on the collected vehicle operation data.

3. The method for estimating the health status of a lithium-ion battery integrating physical information according to claim 1, characterized in that: The step S2 is specifically as follows: S21: analyzing charging data of the electric vehicle, and selecting charging segments that satisfy the conditions that the charging start SOC is less than a first percentage and the charging end SOC is a second percentage as capacity calculation segments; S22: Calculate the charging capacity of the capacity calculation segment by the ampere-hour integration method, and divide it by the SOC change of the capacity calculation segment to obtain a capacity label; S23: Divide the capacity label by the rated capacity of the battery to obtain a health status SOH label; S24: calculating an average cell voltage sequence, a highest cell voltage sequence, and a lowest cell voltage sequence in the capacity calculation segment according to the battery cell voltages; S25: selecting a characteristic start voltage and a characteristic cut-off voltage, and extracting a characteristic extraction segment from the capacity calculation segment according to the average cell voltage sequence, wherein the starting point of the characteristic extraction segment is the moment when the average cell voltage sequence reaches the characteristic start voltage, and the ending point of the characteristic extraction segment is the moment when the average cell voltage sequence reaches the characteristic cut-off voltage; S26: Calculate the cumulative charged power of the feature extraction segment by the ampere-hour integration method as health feature 1, calculate the standard deviation of the power increment sequence of the feature extraction segment by the ampere-hour integration method as health feature 2, calculate the highest single cell voltage at the starting point of the feature extraction segment as feature 3, calculate the lowest single cell voltage at the starting point of the feature extraction segment as feature 4, calculate the highest single cell voltage at the end point of the feature extraction segment as feature 5, calculate the lowest single cell voltage at the end point of the feature extraction segment as feature 6, calculate the range of single cell voltages at the starting point of the feature extraction segment as feature 7, calculate the range of single cell voltages at the end point of the feature extraction segment as feature 8, calculate the average single cell temperature of the feature extraction segment as feature 9, and extract the cumulative mileage of the vehicle as feature 10.

4. The method for estimating the health status of a lithium-ion battery integrating physical information according to claim 1, characterized in that: The step S3 is specifically as follows: S31: A lithium-ion battery SOH estimation model is established based on a multi-layer fully connected neural network. The input of the model is features 1 to 10, and the output of the model is the SOH label. S32: Calculate the partial differential dT of the model output value with respect to feature 9, and define the physical loss 1L1 of the model as the larger of 0 and -dT; S33: Calculate the partial differential dM of the model output value with respect to the feature 10, and define the physical loss 2L2 of the model as the larger of 0 and dM; S34: Calculate the mean absolute percentage error (MAPE) of the model output value and the SOH label, and define MAPE as the estimated loss L3 of the model; S35: Define the total loss L of the model as the weighted sum of L1, L2, and L3; S36: Iteratively train the neural network model according to L and the gradient descent algorithm until L drops to a preset value or there is no longer a downward trend.

5. The method for estimating the health status of a lithium-ion battery integrating physical information according to claim 1, characterized in that: The step S4 is specifically as follows: Step S41: extracting features 1 to 10 from the charging data of the vehicle to be detected; Step S42: Input the features 1 to 10 of the vehicle to be detected into the model, and the model outputs an estimated value of SOH.

Images