A battery state of health evaluation method and a battery state of health evaluation device

Abstract

The application relates to the technical field of battery evaluation, in particular to a battery state of health evaluation method and a battery state of health evaluation device. The battery state of health evaluation method comprises the following steps: collecting complete charging data of a battery charged from a lower limit voltage to an upper limit voltage under different battery state of health, wherein the complete charging data comprises battery voltage, battery temperature, ambient temperature and charging current; constructing a battery health estimation model according to the corresponding relationship between data nodes in the complete charging data and different battery state of health; collecting effective data of a real-time charging trip of a vehicle, and inputting the effective data into the battery health estimation model to obtain a real-time battery state of health value. The application realizes a method for online estimation of the state of health (SOH) of the vehicle and identification of abnormal attenuation of the battery SOH, and the method can realize online estimation of a complete charging curve and the SOH of the battery based on a partial charging interval in a charging process.

Description

Technical Field

[0001] This application relates to the field of battery evaluation technology, specifically to a battery health status evaluation method and a battery health status evaluation device. Background Technology

[0002] As the penetration rate of new energy vehicles in the Chinese automotive market continues to increase, users are also raising their demands for performance, safety, and lifespan. Furthermore, with the gradual retirement of early-stage new energy vehicles, the secondary use of power batteries is becoming a key focus for both the government and industry. To meet these two demands, accurate assessment of the current state of health (SOH) of the power battery is essential. The State of Health (SOH) of a power battery is a crucial indicator for evaluating its performance, reflecting the degree of performance degradation. Accurate SOH estimation and prediction can provide a more reliable range forecast for vehicles, helping to optimize the estimation and strategy for other critical battery states, and ultimately extending battery lifespan.

[0003] In related technologies, the assessment of State of Charge (SOH) in pure electric vehicles mainly focuses on the battery capacity degradation indicator, which is calculated by comparing the battery's current full-charge capacity with the nominal capacity of a new battery. However, in actual vehicle use, calculating SOH in these technologies is quite difficult. This is because SOH requires a complete charging process from the lower voltage limit to the upper voltage limit. However, in actual vehicle use, it is rare to experience a complete charging process from the lower voltage limit to the upper voltage limit. Usually, charging begins at a certain voltage point and is not a complete charging process. It is difficult to accurately calculate the battery's current full-charge capacity using only a partial charging interval, thus making online SOH estimation impossible. Summary of the Invention

[0004] In related technologies, it is difficult to accurately calculate the full charge capacity of a battery in its current state using only a portion of the charging range, which makes it impossible to estimate the battery's health status online.

[0005] In a first aspect, embodiments of this application provide a battery health status assessment method, characterized in that the battery health status assessment method includes:

[0006] Collect complete charging data of the battery from the lower limit voltage to the upper limit voltage under different battery health conditions. The complete charging data includes battery voltage, battery temperature, ambient temperature, and charging current.

[0007] A battery health estimation model is constructed based on the correspondence between data nodes in the complete charging data and different battery health states.

[0008] Collect valid data on the vehicle's real-time charging journey and input the valid data into the battery health estimation model to calculate the real-time battery health status value.

[0009] In conjunction with the first aspect, in one implementation, the step of constructing a battery health estimation model based on the correspondence between data nodes in the complete charging data and different battery health states includes:

[0010] Preprocess and location-encode the complete charging data;

[0011] A time-series regression model is constructed and correlated with preprocessed complete charging data to obtain a battery health estimation model.

[0012] In conjunction with the first aspect, in one implementation, the preprocessing and location encoding of the complete charging data includes:

[0013] The charging process is reconstructed based on the voltage data in the complete charging data to obtain the first complete charging state matrix after reconstruction.

[0014] Position encoding is performed on the first complete charging state matrix.

[0015] In conjunction with the first aspect, in one implementation, the reconstructing of the charging process based on voltage data from the complete charging data includes:

[0016] Calculate the relationship curve between the amount of electricity charged and the time for each charging process in the complete charging data;

[0017] The complete charging data is divided into multiple voltage ranges V, and the charging capacity corresponding to each voltage point in voltage range V is calculated based on the relationship curve between the charged capacity and time. The calculation formula is as follows:

[0018]

[0019] Where Q(t) is the amount of charge added to the battery at time t, and V min V is the lower limit voltage of the voltage range V. max This is the upper limit voltage of the voltage range V;

[0020] The first complete charging state matrix is ​​obtained based on the correspondence between each voltage point in the voltage range V and the battery temperature, ambient temperature, and charging capacity.

[0021] In conjunction with the first aspect, in one embodiment, obtaining the first complete charging state matrix based on the correspondence between each voltage point in the voltage range V and the battery temperature, ambient temperature, and charging capacity further includes:

[0022] If there is no corresponding battery temperature, ambient temperature and charging capacity data for voltage node vi in ​​voltage range V, then the battery temperature, ambient temperature and charging capacity data corresponding to point vi are calculated based on the linear difference between the states of the two voltage nodes adjacent to voltage node vi.

[0023] In conjunction with the first aspect, in one implementation, the position encoding of the first complete charging state matrix includes:

[0024] The first complete charging state matrix is ​​normalized to obtain the second complete charging state matrix;

[0025] The second complete charging state matrix is ​​sliced ​​using a sliding window, and each sliced ​​window is used as the time sequence input and the charging capacity reconstructed after each complete charging process of each window is the target sequence.

[0026] Position encoding is performed on each time-series input and target sequence.

[0027] In conjunction with the first aspect, in one implementation, constructing a time-series regression model and associating the time-series regression model with the preprocessed complete charging data includes:

[0028] A time-series regression model is constructed based on the encoder and decoder;

[0029] Define the loss function for the time series regression model;

[0030] The time-series input of the second complete charging state matrix is ​​used as the input of the time-series regression model, and the model parameter state with the minimum loss is selected as the battery health estimation model using the backpropagation mechanism.

[0031] In conjunction with the first aspect, in one implementation, preprocessing the complete charging data includes:

[0032] The charging process is reconstructed based on the battery state-of-charge data in the complete charging data to obtain the first complete charging state matrix after reconstruction.

[0033] Position encoding is performed on the first complete charging state matrix.

[0034] In conjunction with the first aspect, in one embodiment, the step of collecting valid data on the vehicle's real-time charging journey and transmitting the valid data to a battery health estimation model to calculate a real-time battery health status value includes:

[0035] In the real-time charging data of the vehicle, the charging data in which the battery voltage, charging current, battery temperature, ambient temperature and charging time are all within the preset effective threshold are considered as valid data.

[0036] The effective data is converted into n real-time time series inputs by a sliding window, and the n real-time time series inputs are fed into the battery health estimation model to obtain n battery health state calculation results;

[0037] The effective calculation results y are obtained by filtering n battery health status calculation results. The filtering formula is as follows:

[0038] y∈[Q1-1.5ω*IQR,Q3ω+1.5*ωIQR]

[0039] Where Q1 is the 25th percentile of the n calculation results, Q3 is the 75th percentile of the n calculation results, and IQR is the difference between Q3 and Q1;

[0040] Calculate the average of all valid calculation results in array y, and use the average of the three most recent valid calculation results as the battery health status value.

[0041] Secondly, embodiments of this application provide a battery health status assessment device, which includes:

[0042] The test data collection module is used to collect complete charging data of the battery from the lower limit voltage to the upper limit voltage under different battery health conditions. The complete charging data includes battery voltage, battery temperature, ambient temperature, and charging current.

[0043] The model building module is used to build a battery health estimation model based on the correspondence between data nodes in the complete charging data and different battery health states.

[0044] The evaluation module collects valid data on the vehicle's real-time charging range and feeds this data into the battery health estimation model to calculate the real-time battery health status value.

[0045] The beneficial effects of the technical solutions provided in this application include:

[0046] This invention proposes a data-driven battery SOH estimation method. It establishes a model based on the correspondence between data nodes in the complete charging data and different battery health states. Then, it inputs the battery state transmitted back during the real-time charging process of the vehicle into the model to obtain the corresponding battery health state. This method realizes online estimation of vehicle SOH and identification of abnormal SOH degradation of battery. This method can realize online estimation of battery completion curve and SOH based on a part of the charging process. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the battery health status assessment method of this application;

[0048] Figure 2This is a diagram of the time series regression model architecture for this application;

[0049] Figure 3 This is a schematic diagram illustrating the battery health status assessment process during vehicle operation for this application.

[0050] Figure 4 This is a schematic diagram of the hardware structure of the battery health status assessment device involved in the embodiments of this application. Detailed Implementation

[0051] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0052] In related technologies, it is difficult to accurately calculate the full charge capacity of a battery in its current state using only a portion of the charging range, which makes it impossible to estimate the battery's health status online.

[0053] Firstly, such as Figure 1 and Figure 3 As shown, this application provides a battery health status assessment method, which includes the following steps:

[0054] Step S1: Collect complete charging data of the battery from the lower limit voltage to the upper limit voltage under different battery health conditions. The complete charging data includes battery voltage, battery temperature, ambient temperature, and charging current.

[0055] Specifically, step S1 includes: acquiring complete charging data of the power battery under different battery health conditions on the test bench under the actual charging icon or model (i.e., the process of charging from the lower limit voltage Vmin of the battery to the upper limit voltage Vmax). The complete charging data includes battery voltage V, charging current I, battery temperature T1, ambient temperature T2, and SOC (state of charge, i.e., the state of charge of the battery).

[0056] Step S2: Construct a battery health estimation model based on the correspondence between data nodes in the complete charging data and different battery health states.

[0057] Specifically, step S2 above includes:

[0058] Step a: Preprocess and position-encode the complete charging data.

[0059] Furthermore, this application provides two alternative embodiments for step a:

[0060] In the first embodiment, step a includes: reconstructing the charging process based on the voltage data in the complete charging data to obtain a reconstructed first complete charging state matrix. Then, position encoding is performed on the first complete charging state matrix.

[0061] In the second embodiment, step a includes: reconstructing the charging process based on the SOC data (battery state of charge data) in the complete charging data to obtain a reconstructed first complete charging state matrix. Then, position encoding is performed on the first complete charging state matrix.

[0062] It's worth noting that for batteries undergoing fast charging, the charging current fluctuates significantly depending on battery temperature, SOC, and other conditions, causing voltage rebound. Therefore, voltage-based data reconstruction is not recommended in this situation. For full charging, however, either voltage or SOC-based reconstruction can be used depending on the specific application. If the slow charging current is constant and doesn't fluctuate significantly, preventing voltage rebound, then voltage-based reconstruction is preferable.

[0063] Furthermore, the first implementation of step a above specifically includes:

[0064] Step a1: Calculate the charging capacity.

[0065] Specifically, the relationship curve between the amount of electricity charged and the time during each charging process in the complete charging data is calculated using the following formula:

[0066]

[0067] Where Q(t) is the amount of charge added to the battery at time t, I represents the current at time t, and Δt represents the sampling time step.

[0068] Step a2: Divide the complete charging data into multiple voltage ranges V.

[0069] Specifically, the voltage range [V min V ma x] is divided into N+1 equal parts, i.e., [vmin,vmin+△v,vmin+2*△v,...,vmax], where △V=(Vmax-Vmin) / N. Here, the voltage array after division is denoted as V=[v0,v1,...,vN].

[0070] Step a3: Calculate the charging capacity corresponding to each voltage point in the voltage range V based on the relationship curve between the charged capacity and time. The calculation formula is as follows:

[0071]

[0072] Where Q(t) is the amount of charge added to the battery at time t, and V min V is the lower limit voltage of the voltage range V. max This is the upper limit voltage of the voltage range V.

[0073] Step a4: Obtain the first complete charging state matrix based on the correspondence between each voltage point in voltage range V in the complete charging data and battery temperature, ambient temperature, and charging capacity.

[0074] Specifically, based on the collected complete charging data, a battery temperature T1 = [t0, t1, ..., tN], an ambient temperature A = [a0, a1, ..., aN], and a charging capacity are constructed for each voltage point, V = [v0, v1, ..., vN], resulting in the first complete charging state matrix H = [V; Q; T; A]. Matrix H is:

[0075]

[0076] It is worth noting that in step a4 above, there may be instances where the exact value corresponding to v cannot be found in the complete sampled charging data. i The status at any given time (battery temperature, ambient temperature, and charging capacity data).

[0077] Preferably, when the above-mentioned situation occurs where each state cannot be found, then according to the voltage node v i Calculate the voltage node v by the linear difference between the states of two adjacent voltage nodes. i The corresponding battery temperature, ambient temperature, and charging capacity data are displayed.

[0078] In some specific embodiments, the i-th voltage in the voltage range V is 3.56V, then it is necessary to obtain v i The charging capacity q of the battery at 3.56V. i Battery temperature t i and ambient temperature a i However, at this time, because there was no v during the original data sampling process... i =3.56V, therefore the corresponding charging capacity q cannot be directly obtained. i Battery temperature t i and ambient temperature a i At this point, it is necessary to perform linear interpolation using the two states corresponding to 3.55V and 3.57V closest to 3.56V to obtain the voltage node v. i Corresponding charging capacity q i Battery temperature t i and ambient temperature a i .

[0079] Step a5: Normalize each column of the first complete charging state matrix H according to the following formula to obtain the normalized second complete charging state matrix S. The formula is as follows:

[0080]

[0081] Among them, h ij For each element in the first complete charging state matrix, u j σ is the average value of the column containing the element in the first complete charging state matrix. j This represents the standard deviation of the column containing the element in the first complete charge state matrix. It's worth noting that normalizing the first complete charge state matrix H simplifies subsequent calculations.

[0082] Step a6: Use a sliding window to slice the second complete charging state matrix, and construct each sliced ​​window as the time-series input X and the charging capacity reconstructed for each complete charging process of each window as the target sequence Y.

[0083] Specifically, step a6 includes:

[0084] The second complete charging state matrix S = [V; Q; T; A] is a tensor of shape (N+1, 4). S is divided into segments by a sliding window of fixed width L, and each segment is denoted as x. i ,in,

[0085] Each x i Let X be a tensor of shape (L, 4). The total number of partitioned windows is denoted as X, where X is a tensor of shape (N-L+2, L, 4). Each partitioned window is shown in the following matrix form:

[0086] (where i = 0, 1, 2, ..., N-L+1)

[0087] The battery capacity q in the second column is processed by subtracting q1 from each value of the battery capacity q in the second column, as shown in submatrix 2 below.

[0088]

[0089] (where i = 0, 1, 2, ..., N-L+1)

[0090] Wherein, each constructed time-series input x i Its corresponding training target sequence y i The reconstructed battery capacity for each complete charging cycle.

[0091] Step a7: Input x for each timing sequence according to the formula. iIts corresponding training target sequence y i Perform position encoding.

[0092] Specifically, the encoding formula in step a7 above is as follows:

[0093]

[0094] Where k represents the k-th row of matrix x, and i represents the i-th column of matrix x. The timing input x... i Add the corresponding encoding position to the training target sequence y.

[0095] Step b: Construct a time-series regression model and correlate it with the preprocessed complete charging data to obtain a battery health estimation model.

[0096] Specifically, step b above includes:

[0097] Step b1: Construct a temporal regression model based on the encoder and decoder.

[0098] Specifically, an encoder-decoder architecture is established as follows: Figure 2 The temporal regression model shown. The encoder layer input is the temporal input x. i Timing input x i After n layers of encoding operations, the result e is... i The inputs to the decoding layer are [xi; p0] and the result e from the encoder layer. i Where p0 is a zero-based evidence of shape (N+1,4), pred represents the length of the sequence to be predicted, and the result d is obtained after m decoding operations. i The prediction result is passed to a fully connected layer.

[0099] Preferably, the encoder described above is composed of multiple superimposed encoding operations. In each encoding operation, the timing input x... i First, a self-attention mechanism is performed, followed by a 1D convolutional network. The decoder described above is composed of multiple decoding operations stacked together. In each decoding operation, x... i First, perform self-attention operation, and then... (The sentence is incomplete and requires more context to translate accurately.) i Perform cross-attention operations.

[0100] Step b2: Define the loss function for the time series regression model.

[0101] Specifically, MSE, as the loss function of the model, is formulated as follows:

[0102]

[0103] Step b3: Input x with the timing of the second complete charging state matrix S. i As input to the time-series regression model, the backpropagation mechanism is used to select the model parameter state with the minimum loss as the battery health estimation model.

[0104] In some preferred embodiments, step b3 above further includes: optimizing and correcting the battery health estimation model. The correction process includes: after obtaining the complete charging process status of the vehicle during vehicle maintenance, the data is input into the model to evaluate the model's loss. When the loss is higher than the set value lmax, historically obtained data is brought into the model for training and fine-tuning.

[0105] It is worth noting that the historical data includes the test data collected in step S1 or the charging data during vehicle operation. Meanwhile, the battery health estimation model is deployed in the cloud, which can receive real-time data from the vehicle.

[0106] Step S3: Collect valid data on the vehicle's real-time charging journey and input the valid data into the battery health estimation model to calculate the real-time battery health status value.

[0107] Specifically, step S3 above includes:

[0108] Step S3a: Select the charging data in the real-time vehicle charging data where the battery voltage, charging current, battery temperature, ambient temperature, and charging time are all within the preset effective thresholds during the charging process as valid data.

[0109] It's worth noting that testing the State of Harmony (SOH) of a battery under laboratory conditions is relatively easy, but in real-world applications, battery operating conditions can vary greatly, such as temperature and charging rate. These factors can affect battery capacity and performance, leading to discrepancies between laboratory test results and the SOH observed in actual use. Therefore, in step S3a above, the real-time vehicle charging data is filtered before being sent back to the cloud for processing and calculation.

[0110] Step S3b: After inputting the valid data into the cloud, the valid data is converted into n real-time time series inputs through a sliding window, and the n real-time time series inputs are input into the battery health estimation model to obtain n battery health status calculation results.

[0111] Specifically, step S3b above includes: the cloud reconstructs the valid data uploaded by the vehicle using the resampling method described in step S2 above. Then, it converts this data into n inputs identical to xi using a sliding window method (n is related to the length of the data collected during charging), and substitutes these into the model to obtain n battery health state calculation results. The average of the m effective results is then calculated to obtain the current battery health state value of the vehicle. The set of effective calculation results y is defined as y∈[Q1ω-ω1.5ω*IQR,Q3ω+1.5*IQR].

[0112] Where Q1 is the 25th percentile of the n results, Q3 is the 75th percentile of the n results, and IQR = Q3 - Q1. Finally, the average of the battery health status calculation results obtained from the vehicle's most recent three calculations is taken as the final real-time battery health status calculation result from the cloud and fed back to the vehicle.

[0113] Secondly, embodiments of this application also provide a battery health status assessment device.

[0114] In one embodiment, the battery health status assessment device includes: an experimental data collection module, a model building module, and an assessment module; wherein,

[0115] The test data collection module is used to collect complete charging data of the battery from the lower limit voltage to the upper limit voltage under different battery health states. The complete charging data includes battery voltage, battery temperature, ambient temperature, and charging current. The model building module is used to build a battery health estimation model based on the correspondence between data nodes in the complete charging data and different battery health states. The evaluation module is used to collect valid data of the vehicle's real-time charging journey and input the valid data into the battery health estimation model to calculate the real-time battery health state value.

[0116] The functions of each module in the battery health status assessment device correspond to the steps in the battery health status assessment method embodiment, and their functions and implementation processes will not be described in detail here.

[0117] Thirdly, embodiments of this application provide a battery health status assessment device, which can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0118] Reference Figure 4 , Figure 4 This is a schematic diagram of the hardware structure of the battery health status assessment device involved in the embodiments of this application. In the embodiments of this application, the battery health status assessment device may include a processor, a memory, a communication interface, and a communication bus.

[0119] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0120] Communication interfaces include input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting devices within the AAAA device, as well as interfaces used for interconnecting the AAAA device with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0121] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0122] The processor can be a general-purpose processor, which can call the AAAA program stored in memory and execute the AAAA method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the AAAA program is called can be referred to in the various embodiments of the AAAA method of this application, and will not be repeated here.

[0123] Those skilled in the art will understand that Figure 4 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0124] Fourthly, embodiments of this application also provide a readable storage medium.

[0125] This application has a readable storage medium storing a battery health status assessment program, wherein when the battery health status assessment program is executed by a processor, it implements the steps of the battery health status assessment method as described above.

[0126] The method implemented when the battery health status assessment procedure is executed can be referred to in various embodiments of the battery health status assessment method of this application, and will not be repeated here.

[0127] In summary, this invention proposes a data-driven battery SOH estimation method. It establishes a model based on the correspondence between data nodes in the complete charging data and different battery health states. Then, it inputs the battery state transmitted back during real-time vehicle charging into the model to obtain the corresponding battery health state. This enables online estimation of vehicle SOH and identification of abnormal battery SOH degradation. This method can achieve online estimation of the battery's charging curve and SOH based on a portion of the charging process.

[0128] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0130] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0131] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0132] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0133] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0134] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method of battery state of health estimation, characterized by, The battery health state evaluation method comprises: Collecting complete charging data of the battery charged from a lower limit voltage to an upper limit voltage under different battery health states, the complete charging data comprising battery voltage, battery temperature, ambient temperature and charging current; Building a battery health estimation model according to the correspondence between data nodes in the complete charging data and different battery health states; Collecting effective data of a real-time charging trip of the vehicle, and feeding the effective data into the battery health estimation model to obtain a real-time battery health state value; The battery health estimation model is built according to the correspondence between the data nodes in the complete charging data and the different battery health states, comprising: preprocessing and position coding of the complete charging data; building a time series regression model, and associating the time series regression model with the preprocessed complete charging data to obtain the battery health estimation model; The time series regression model is built and associated with the preprocessed complete charging data, comprising: building a time series regression model based on an encoder and a decoder; defining a loss function of the time series regression model; taking the time series input of the second complete charging state matrix as the input of the time series regression model, and using the back propagation mechanism to screen out the model parameter state with the minimum loss as the battery health estimation model.

2. The battery state of health assessment method of claim 1, wherein, The preprocessing and position coding of the complete charging data comprise: Reconstructing the charging process based on the voltage data in the complete charging data to obtain a first complete charging state matrix; Position coding the first complete charging state matrix.

3. The battery state of health assessment method of claim 2, wherein, The reconstruction of the charging process based on the voltage data in the complete charging data comprises: Calculating the relationship curve between the charging capacity and time of each charging process in the complete charging data; Dividing the complete charging data into a plurality of voltage intervals V, and calculating the charging capacity corresponding to each voltage point in the voltage interval V according to the relationship curve between the charging capacity and time, the calculation formula being as follows: wherein Q(t) is the amount of electricity charged into the battery at time t, V min is the lower limit voltage of the voltage interval V, V max is the upper limit voltage of the voltage interval V; Obtaining the first complete charging state matrix according to the correspondence between each voltage point in the voltage interval V and the battery temperature, the ambient temperature and the charging capacity.

4. The battery state of health assessment method of claim 3, wherein, The obtaining of the first complete charging state matrix according to the correspondence between each voltage point in the voltage interval V and the battery temperature, the ambient temperature and the charging capacity further comprises: When there is no corresponding battery temperature, ambient temperature and charging capacity data for the voltage node vi in the voltage interval V, the battery temperature, ambient temperature and charging capacity data corresponding to the voltage node vi are calculated according to the linear difference value of the corresponding states of the two voltage nodes adjacent to the voltage node vi.

5. The battery state of health assessment method of claim 2, wherein, The position coding of the first complete charging state matrix comprises: Normalizing the first complete charging state matrix to obtain a second complete charging state matrix; Slicing the second complete charging state matrix using a sliding window, and building each window after slicing into a time series input and a target sequence of the charging capacity after reconstruction of each complete charging process in the window; Position coding each time series input and target sequence.

6. The battery state of health assessment method of claim 5, wherein, The collection of the effective data of the real-time charging trip of the vehicle, and the feeding of the effective data into the battery health estimation model to obtain the real-time battery health state value comprise: The charging data in which the battery voltage, the charging current, the battery temperature, the ambient temperature and the charging time in the charging process are all within the preset effective threshold are taken as effective data in real-time charging data of a vehicle; The effective data is converted into n real-time time series inputs through a sliding window, and the n real-time time series inputs are input into the battery health estimation model to obtain n battery health state calculation results; The n battery health state calculation results are screened to obtain an array y of effective calculation results, and the screening formula is: wherein Q1 is the 25% quantile of the n calculation results, Q3 is the 75% quantile of the n calculation results, and IQR is the difference between Q3 and Q1; The average value of all the effective calculation results in the array y is calculated, and the average value of the last three effective calculation results is taken as the battery health state value.

7. The battery state of health assessment method of claim 1, wherein, The preprocessing and position coding of the complete charging data include: The charging process is reconstructed based on the battery state of charge data in the complete charging data to obtain a first reconstructed complete charging state matrix; The first reconstructed complete charging state matrix is position coded.

8. A battery state of health assessment apparatus characterized by comprising: The battery health state evaluation device includes: A test data collection module is configured to collect complete charging data of a battery charged from a lower limit voltage to an upper limit voltage under different battery health states, wherein the complete charging data includes battery voltage, battery temperature, ambient temperature and charging current; A model construction module is configured to construct a battery health estimation model according to the correspondence between data nodes in the complete charging data and different battery health states; The battery health estimation model is constructed according to the correspondence between the data nodes in the complete charging data and the different battery health states, including: preprocessing and position coding of the complete charging data; constructing a time series regression model and associating the time series regression model with the preprocessed complete charging data to obtain the battery health estimation model; The time series regression model is constructed and associated with the preprocessed complete charging data, including: constructing a time series regression model based on an encoder and a decoder; defining a loss function of the time series regression model; taking the time series input of the second complete charging state matrix as the input of the time series regression model, and using a back propagation mechanism to screen out the model parameter state with the minimum loss as the battery health estimation model; An evaluation module is configured to collect effective data of a real-time charging trip of a vehicle and input the effective data into the battery health estimation model to obtain a real-time battery health state value.

Images