Battery state of health diagnosis method and device, electronic equipment and storage medium

Abstract

The application provides a battery health state diagnosis method and device, electronic equipment and storage medium, and relates to the technical field of battery detection. The method comprises the following steps: obtaining battery data for data preprocessing to obtain a standard data set; a state of charge-open circuit voltage curve of a battery cell is constructed; the single-point health degree of each battery cell in a single charging segment is estimated according to the standard data set and the state of charge-open circuit voltage curve; a health degree sequence is obtained according to the single-point health degree, and the health degree sequence is subjected to data enhancement; the health degree sequence after data enhancement is subjected to double-layer filtering to obtain an optimal estimated health degree sequence; and the health state of the battery is diagnosed according to the optimal estimated health degree sequence. The application realizes an adaptive updating mode of the health state degree, can generalize the health state estimation of the whole life cycle under real working conditions, can generalize different types of batteries at the same time, and effectively improves the accuracy, stability and generalization of the model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery detection, and in particular to a battery health state diagnosis method and device, an electronic device and a storage medium. BACKGROUND

[0002] With the rapid development of new energy vehicles, lithium ion batteries are widely used due to their high energy density, low self-discharge rate, long cycle life, energy saving and environmental protection, and other outstanding advantages. With the calendar and cycle use of lithium ion batteries, lithium ion batteries are prone to problems such as capacity reduction, performance reduction, and deterioration of safety and stability. Therefore, it is of great engineering significance to carry out research on the health evaluation and diagnosis of lithium ion power batteries. However, once the lithium ion battery is installed on the vehicle, the maximum available capacity of the battery cannot be directly measured, and there are various complex chemical reactions and physical changes inside the lithium ion battery, making it more complex to accurately quantify the health state.

[0003] Under actual working conditions, the existing method mainly collects real-time information such as voltage, current and temperature of the battery through the battery management system at the vehicle end to make predictions. By establishing an equivalent circuit model or a life prediction model of the battery, the health degradation trend of the battery under different rates, temperatures and discharge depths is predicted, so as to estimate the battery health in different periods. However, this method is strongly dependent on the battery model. As the aging degree of the battery increases, the parameters of the battery model change in nature, and the complex electrochemical changes cannot be accurately represented by the battery model. In particular, in the middle and later stages, adaptive model updating and iteration cannot be performed, and the accuracy and stability of this prediction method are greatly reduced. The battery aging model trained in the laboratory conditions cannot be generalized to real working conditions, and its prediction accuracy and stability cannot achieve the expected effect in the whole life cycle. SUMMARY

[0004] In view of the above shortcomings of the prior art, the present application provides a battery health state diagnosis method, device, electronic device and storage medium, which effectively solves the problem of low accuracy and stability of the existing prediction method.

[0005] In a first aspect, the present application provides a battery health state diagnosis method, which comprises:

[0006] Obtaining battery data for data preprocessing to obtain a standard data set;

[0007] Constructing a state of charge-open circuit voltage curve of the battery cell;

[0008] Estimating the single-point health degree of each cell in a single charging segment according to the standard data set and the state of charge-open circuit voltage curve;

[0009] obtaining a health degree sequence according to the single-point health degree;

[0010] performing double-layer filtering on the health degree sequence after data enhancement to obtain an optimal estimated health degree sequence;

[0011] diagnosing the health state of the battery according to the optimal estimated health degree sequence.

[0012] Further, after obtaining the optimal estimated health degree, the method further comprises:

[0013] reconstructing parameters of the state of charge-open circuit voltage curve according to the optimal estimated health degree sequence to obtain a state of charge-open circuit voltage curve under optimal health degree.

[0014] Further, the estimating the single-point health degree of each battery cell in each charging segment according to the state of charge-open circuit voltage curve comprises:

[0015] obtaining a charging segment in which a static state exists at the beginning and end of charging in a charging working condition;

[0016] inputting a first voltage of each battery cell at the beginning of charging in the charging segment into the state of charge-open circuit voltage curve to obtain a first state of charge of each battery cell;

[0017] inputting a second voltage of each battery cell at the end of charging in the charging segment into the state of charge-open circuit voltage curve to obtain a second state of charge of each battery cell;

[0018] cumulatively integrating the current of the charging segment to obtain a real-time charging capacity of the charging segment;

[0019] calculating the single-point health degree of each battery cell according to the first state of charge, the second state of charge and the real-time charging capacity.

[0020] Further, the obtaining a health degree sequence according to the single-point health degree and performing data enhancement on the health degree sequence comprise:

[0021] constructing a variational autoencoder model;

[0022] performing time sequence calculation according to the single-point health degree to obtain a time sequence health degree matrix;

[0023] training the variational autoencoder model according to the time sequence health degree matrix to obtain an optimal variational autoencoder model;

[0024] constructing the health degree sequence according to the single-point health degree, and inputting the health degree sequence into the optimal variational autoencoder model for data enhancement.

[0025] Further, the data-enhanced health degree sequence is subjected to double-layer filtering to obtain an optimal estimated health degree sequence, including:

[0026] A Gaussian process regression model is constructed;

[0027] The global sequence points of the data-enhanced health degree sequence are filtered by using the Gaussian process regression model to obtain sequence estimation values;

[0028] Noise points in the health degree sequence are removed according to the sequence estimation values to obtain a non-noise point sequence;

[0029] The non-noise point sequence is subjected to local filtering by using the Gaussian process regression model to obtain the optimal estimated health degree sequence.

[0030] Further, the health state of the battery is diagnosed according to the optimal estimated health degree sequence, including:

[0031] The attenuation rate of each cell is calculated according to the optimal estimated health degree sequence;

[0032] The health degree outlier of each cell is calculated according to the optimal estimated health degree sequence;

[0033] The health state of the battery is diagnosed according to the attenuation rate and / or the health degree outlier.

[0034] Further, the state of charge-open circuit voltage curve is parameter reconstructed according to the optimal estimated health degree sequence to obtain a state of charge-open circuit voltage curve under an optimal health degree, including:

[0035] The state of charge of each cell is calibrated according to the optimal estimated health degree sequence;

[0036] The open circuit voltage value of each cell is calculated according to the calibrated state of charge and the state of charge-open circuit voltage curve to obtain a state of charge-open circuit voltage matrix;

[0037] The state of charge-open circuit voltage matrix is aggregated to obtain a state of charge-open circuit voltage curve under an optimal health degree.

[0038] In a second aspect, the present application provides a battery health state diagnosis device, including:

[0039] A data acquisition module is configured to acquire battery data, perform data preprocessing, and obtain a standard data set;

[0040] A curve construction module is configured to construct a state of charge-open circuit voltage curve of a battery cell;

[0041] a data estimation module configured to estimate a single-point health degree of each battery cell in a single charging segment according to the standard data set and the state of charge-open circuit voltage curve;

[0042] a data enhancement module configured to obtain a health degree sequence according to the single-point health degree and perform data enhancement on the health degree sequence;

[0043] a data filtering module configured to perform double-layer filtering on the health degree sequence after data enhancement to obtain an optimal estimated health degree sequence;

[0044] a state diagnosis module configured to diagnose the health state of the battery according to the optimal estimated health degree sequence.

[0045] In a third aspect, the present application provides an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the battery health state diagnosis method according to the first aspect of the present application.

[0046] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program, and the computer program is executable on a processor to implement the battery health state diagnosis method according to the first aspect of the present application.

[0047] The battery health state diagnosis method, device, electronic device, and storage medium provided by the present application adopt the method of estimating the optimal health state degree to complete dynamic reconstruction of the state of charge-open circuit voltage curve, realize self-update of model parameters, and significantly improve the model generalization effect and stability under real working conditions through the data enhancement method for sparse data under real working conditions. Based on the double-layer filtering mode, the adaptive update mode of the health state degree is realized, which can generalize the health stable estimation of the whole life cycle under real working conditions and generalize the batteries of different battery types, effectively improving the accuracy, stability, and generalization of the model. The health diagnosis of the battery cells or modules in the battery pack can be realized, and the health state diagnosis at the cell level also provides strong support for the maintenance and abnormal diagnosis of the battery cells or modules. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0049] Figure 1 is the first schematic diagram of the battery health state diagnosis method flow provided by the embodiments of the present application;

[0050] Figure 2 is a second schematic diagram of a battery health state diagnosis method provided by an embodiment of the application;

[0051] Figure 3 is a third schematic diagram of a battery health state diagnosis method provided by an embodiment of the application;

[0052] Figure 4 is a schematic diagram of a variational autoencoder model in an embodiment of the application;

[0053] Figure 5 is an original health degree sequence in an embodiment of the application;

[0054] Figure 6 is a new health degree sequence generated by a variational autoencoder model in an embodiment of the application;

[0055] Figure 7 is a fourth schematic diagram of a battery health state diagnosis method provided by an embodiment of the application;

[0056] Figure 8 is a schematic diagram of a result of data filtering in an embodiment of the application;

[0057] Figure 9 is a fifth schematic diagram of a battery health state diagnosis method provided by an embodiment of the application;

[0058] Figure 10 is a sixth schematic diagram of a battery health state diagnosis method provided by an embodiment of the application;

[0059] Figure 11 is a seventh schematic diagram of a battery health state diagnosis method provided by an embodiment of the application;

[0060] Figure 12 is a schematic diagram of state of charge-open circuit voltage curve reconstruction in an embodiment of the application;

[0061] Figure 13 is a schematic diagram of a result of a battery health state diagnosis device provided by an embodiment of the application;

[0062] Figure 14 is a schematic diagram of a structure of an electronic device provided by an embodiment of the application.

[0063] Main element symbol explanation:

[0064] 800, battery health state diagnosis device; 810, data acquisition module; 820, curve construction module; 830, data estimation module; 840, data enhancement module; 850, data filtering module; 860, state diagnosis module; 900, electronic device; 910, processor; 920, communication interface; 930, memory; 940, communication bus. DETAILED DESCRIPTION

[0065] In order to make the objects, technical solutions and advantages of the present application clearer, the following will further clearly and completely describe the technical solutions of the present application with reference to the accompanying drawings of the embodiments of the present application. It should be explained that the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work under the premise should belong to the protection scope of the present application.

[0066] In addition, the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more than two, unless otherwise explicitly and specifically limited.

[0067] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the present application is only for the purpose of describing specific embodiments and is not intended to limit the present application. The term "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0068] The battery state of health (SOH) is used to evaluate the maximum intrinsic capacity that the battery can discharge at the current state. The SOH prediction at the vehicle end is mainly performed by the battery management system (BMS) to collect real-time information such as voltage, current and temperature of the battery, to predict the SOH decay trend of the battery at different rates, temperatures and discharge depths by establishing an equivalent circuit model or a life prediction model of the battery, so as to estimate the battery SOH in different periods. However, this method strongly depends on the battery model. As the aging degree of the battery increases, the parameters of the battery model change in nature, and the complex electrochemical changes cannot be accurately characterized by the battery model, especially in the middle and later stages, the model cannot be updated and iterated adaptively, and the accuracy and stability of the prediction method are greatly reduced. At the same time, limited by the limited computing resources of the local embedded system, the complexity of the battery model is greatly reduced, and the precise estimation of the full life cycle of different batteries cannot be realized, so that the battery cannot be accurately diagnosed and evaluated.

[0069] Therefore, the related skilled person uploads the vehicle-end battery data to the cloud, and the cloud analyzes and predicts the SOH based on the battery operation data. The existing method mainly establishes the mapping relationship of the battery with the cycle and calendar aging factors based on the aging data of the laboratory battery, so as to realize the prediction of the SOH of the battery according to the aging factors. However, this is an ideal working condition prediction, and the laboratory working condition (constant current charging, and full charging and discharging state) is extremely simple, and cannot be generalized to the SOH estimation under the real working condition. Under the real working condition, the current, temperature and voltage of the battery are extremely variable, and the battery operation data is relatively sparse, because the vehicle-end may run or be stagnant for a long time, the sampling frequency of the real vehicle data is greatly reduced, and the sampling period is relatively long, which cannot be compared with the stable and single working condition of the laboratory. Therefore, the same aging factors as the laboratory cannot be extracted under the real working condition, so that the battery aging model trained under the laboratory working condition cannot be generalized to the real working condition, and the prediction accuracy and stability cannot achieve the expected effect in the whole life cycle.

[0070] Embodiment 1

[0071] The embodiment of the application provides a battery health state diagnosis method, which effectively solves the problem that the accuracy and stability of the existing prediction method cannot meet the whole life cycle prediction requirements of the battery. Figure 1 is the first schematic diagram of the battery health state diagnosis method provided by the embodiment of the application, as shown in the figure, the method comprises the following steps: Figure 1

[0072] S100, obtain battery data for data preprocessing to obtain a standard data set.

[0073] In the embodiment of the application, the health degree of the lithium ion battery of the vehicle is predicted based on the cloud scenario. Under the real working condition, the power or energy storage system of the vehicle will continuously generate new battery data during operation, the BMS collects, monitors and manages the voltage, current and temperature of the battery, and uploads the data to the cloud system in real time. The cloud system and the vehicle-end comply with the GB32960 protocol (a national standard protocol for transmitting data from the local end to the cloud), and the cloud system obtains and stores the battery data in real time for big data analysis.

[0074] ​After the cloud obtains the battery data, the battery cell voltage data, temperature data, current data and charging and discharging states are mainly analyzed, and abnormal values are cleaned. In the embodiment of the application, the battery cell voltage data can be a voltage value generated by each battery cell at each time, and a series system composed of multiple battery cells forms a voltage list of all voltage data, and similarly, a temperature list, a current list and a charging and discharging state list can be formed. Abnormal values include but are not limited to illegal values, jump values, sampling abnormal values, mutation values, signal jitter and other abnormal types of values. For illegal values (for example, the reasonable range of battery cell voltage is within 1.8V-4.5V, and values outside the range are considered illegal values), direct elimination is performed. For jump values, mutation values and signal jitter values, forward filling, average filling or weighted filling and other treatments are performed. For sampling abnormal values, the event range is eliminated.

[0075] After data cleaning, the working conditions are classified according to the vehicle state and the charging state, mainly divided into charging conditions, static conditions and discharging conditions. Specifically, if the battery is in a charging state, it is identified as a charging condition, if the battery is in an uncharged state and in a discharging state, it is identified as a discharging condition, and if the battery is in an uncharged and uncharged state, it is identified as a static condition.

[0076] S200, a state of charge-open circuit voltage curve of the battery cell is constructed.

[0077] When the lithium ion battery is in an open circuit state, that is, no current passes through the external circuit, at this time the chemical reaction inside the battery will tend to be in a balanced state, and the potential difference between the positive and negative electrodes reaches a stable value, which is the open circuit voltage (OCV). The state of charge (SOC) of the lithium ion battery and the OCV have a certain mapping relationship, and this relationship is the state of charge-open circuit voltage curve (SOC-OCV curve). The curve is used to verify the current SOC of the battery cell. Generally, the OCV of the lithium battery increases with the increase of the SOC, showing a specific curve relationship, and the real SOC of the battery cell can be obtained by the OCV collected by the BMS.

[0078] In the embodiment of the application, the SOC and OCV of the initial battery cell can be measured in a laboratory environment, and the SOC-OCV curve is calibrated according to the measured SOC and OCV to obtain the initial SOC-OCV curve.

[0079] S300, a single-point health degree of each battery cell in a single charging segment is estimated according to the standard data set and the state of charge-open circuit voltage curve.

[0080] In the embodiment of the application, the charging condition under the real working condition is selected, and the single SOH estimation is performed on the charging segment of the charging start and the charging end under the static condition.Figure 2 is a second schematic diagram of a battery state of health diagnosis method provided by an embodiment of the application, as shown in Figure 2 The calculation of the single-point health degree includes the following steps:

[0081] S310, a charging segment in which a static state exists at the beginning and end of charging in the charging condition is obtained.

[0082] In the embodiment of the application, the charging segment in which the static state exists at the beginning of charging for a duration exceeding a first time threshold and the static state exists at the end of charging for a duration exceeding a second time threshold is selected, and the first time threshold and the second time threshold can be set according to actual conditions.

[0083] S320, the first voltage of each cell at the beginning of charging in the charging segment is input into the state of charge-open circuit voltage curve to obtain the first state of charge of each cell.

[0084] When the static state exists at the beginning of charging for a duration exceeding the first time threshold in the charging condition, the first voltage of each cell at the beginning of charging in the charging segment is obtained to form a first voltage list, and the first voltage list is input into the SOC-OCV curve to obtain the first state of charge of each cell and form a first state of charge list.

[0085] S330, the second voltage of each cell at the end of charging in the charging segment is input into the state of charge-open circuit voltage curve to obtain the second state of charge of each cell.

[0086] When the static state exists at the end of charging for a duration exceeding the second time threshold, the second voltage of each cell at the end of charging in the charging segment is obtained to form a second voltage list, and the second voltage list is input into the SOC-OCV curve to obtain the second state of charge of each cell and form a second state of charge list.

[0087] S340, the current of the charging segment is accumulated for ampere-hour integration to obtain the real-time charging capacity of the charging segment.

[0088] The ampere-hour integration method integrates the charging current of the charging segment, and the real-time charging capacity of the charging segment can be obtained through the integration of current and time. The basic principle is that the change of the electric quantity of the battery is proportional to the integral of the current through the battery.

[0089] S350, the single-point health degree of each cell is calculated and obtained according to the first state of charge, the second state of charge and the real-time charging capacity.

[0090] In the embodiment of the application, the single-point health degree calculation formula of each cell is as follows:

[0091]

[0092] In the above formula, SOH1 represents a single-point health degree, C E represents a real-time charging capacity, SOC1 represents a first state of charge, SOC2 represents a second state of charge, C N represents a rated capacity calibrated at the factory of the battery.

[0093] The single-point health degree of each battery cell is calculated to form a single-point health degree matrix. For example, if a battery includes 240 battery cells, the single-point health degree matrix has 240 single-point health degrees.

[0094] S400, obtaining a health degree sequence according to the single-point health degree, and performing data enhancement on the health degree sequence.

[0095] A certain trigger condition is required to estimate the single-point health degree, that is, there is a static working condition at the front and rear ends of the charging segment, and the charging depth exceeds a threshold, wherein the charging depth is SOC2-SOC1. In real working conditions, it is impossible to calculate every day under the trigger condition, for example, only 10 estimations may be triggered in a month, so there are a large number of sparse data. Therefore, missing values need to be captured according to a large number of single-point health degree trends, and missing points are automatically generated so as to subsequently calculate the optimal estimation battery health degree.

[0096] Figure 3 is a third schematic diagram of a battery health state diagnosis method provided by an embodiment of the present application, as shown in Figure 3 The data enhancement specifically includes the following steps:

[0097] S410, constructing a variational autoencoder model.

[0098] In the embodiment of the present application, a generative neural network model, that is, a variational autoencoder model, Figure 4 is an architecture schematic diagram of the variational autoencoder model in the embodiment of the present application, as shown in Figure 4 The architecture of the variational autoencoder model mainly includes an encoder and a decoder. The encoder is responsible for compressing the input data into a representation of a latent space, and the decoder reconstructs the original data from the representation.

[0099] The loss function of the variational autoencoder model is composed of two parts, that is, a reconstruction loss Loss_rec and a KL divergence loss Loss_KL, wherein:

[0100] The purpose of the reconstruction loss is to make the reconstructed data as close as possible to the original input data, so as to ensure that the model can effectively learn the features and distribution of the data, so as to accurately restore the data during decoding. In the embodiment of the present application, mean square error MSE is adopted for representation, and the function expression of the reconstruction loss is as follows:

[0101]

[0102] In the above formula, Loss rec represents the function of reconstruction loss, n represents the number of samples, x i represents the time series window data of the i-th input, represents the average value of the input time series window data.

[0103] The KL divergence loss is used to measure the difference between two probability distributions, i.e. the difference between the distribution of the latent variable Z and the prior distribution N(0, 1). The role of the KL divergence loss is to make the distribution of the latent variable Z output by the encoder as close as possible to the prior distribution, thereby ensuring the continuity and normality of the latent space, making it more reasonable and effective to sample and generate new data in the latent space, effectively preventing overfitting. The function expression of the KL divergence loss is as follows:

[0104]

[0105] In the above formula, Loss KL represents the function of KL divergence loss, n represents the number of samples, μ i represents the average value of the i-th sample, σ i represents the standard deviation of the i-th sample.

[0106] The total loss function of the model is the weighted sum of the reconstruction loss and the KL divergence loss, and the expression of the total loss function of the model is as follows:

[0107] Loss model = Loss rec + α·Loss EL

[0108] In the above formula, Loss model represents the total loss function of the model, α represents the hyperparameter, used to measure the importance of Loss rec and Loss KL .

[0109] By minimizing this total loss function, the variational autoencoder model can simultaneously generate samples similar to the original data and learn the distribution of the latent space.

[0110] S420, time series calculation is performed according to the single-point health degree to obtain a time series health degree matrix.

[0111] The single-point health degree in a certain period of time is time series calculated to obtain a time series health degree matrix, which can be a month or several months.

[0112] S430, training a variational autoencoder model according to the time series health degree matrix to obtain an optimal variational autoencoder model.

[0113] The time sequence health degree matrix is converted into a time sequence data curve, a large amount of time sequence data curves are input into a full connection neural network to encode input features, a latent space distribution of the input is learned by using the neural network, and the input features are resampled to obtain latent variables Z, then the resampled latent variables Z are decoded, a loss function is used to minimize the distribution of the output and the input, and finally an optimal variational autoencoder model is obtained. The training process of the variational autoencoder model does not require additional labels, so there is no additional label acquisition cost.

[0114] S440, according to the single-point health degree, a health degree sequence is constructed, and the health degree sequence is input into the optimal variational autoencoder model for data enhancement.

[0115] The single-point health degree under the real working condition is acquired, and a health degree sequence is constructed, which is usually a sequence with data missing points. The health degree sequence is input into the optimal variational autoencoder model to estimate the data generation probability near the data missing point, and according to the long sequence rule learned by the optimal variational autoencoder model, a new health degree sequence is regenerated.

[0116] Figure 5 is the original health degree sequence in the embodiment of the application, Figure 6 is a new health degree sequence generated by the variational autoencoder model in the embodiment of the application, according to Figure 5 and Figure 6 analysis, the new health degree sequence has more data and a better SOH decay trend, which greatly improves the generalization performance and stability of the optimal estimation of SOH.

[0117] S500, the health degree sequence after data enhancement is subjected to double-layer filtering to obtain an optimal estimation health degree sequence.

[0118] As shown in Figure 6 , the health degree sequence after data enhancement is affected by different working condition environments and data sampling periods, and presents different degrees of jitter and noise, so the new health degree sequence needs to be filtered to obtain an optimal estimation health degree sequence. Figure 7 is the fourth schematic diagram of the battery health state diagnosis method provided by the embodiment of the application, as shown in Figure 7 , the data filtering specifically includes the following steps:

[0119] S510, a Gaussian process regression model is constructed.

[0120] Gaussian Process Regression (GPR) is a non-parametric regression method based on Bayesian statistics, which can model the potential distribution of data by Gaussian Process and give the prediction value and the corresponding uncertainty estimate. Gaussian Process is a random process in which any finite number of random variables obeys joint Gaussian distribution, and a Gaussian Process is completely defined by a mean function and a kernel function, that is:

[0121] f(x) ~ GP(m(x), k(x, x'))

[0122] In the above formula, f(x) represents the optimal estimated health degree, m(x) represents the mean function, and k(x, x') represents the kernel function, which is used to describe the similarity between input sequences.

[0123] In the embodiment of the present application, the time series SOH matrix of a given window is to obtain the optimal estimated health degree, wherein x i and y i represent the corresponding time series data, and the threshold of the window can be set according to the actual situation. The kernel function uses Markov kernel function, and its expression is as follows:

[0124]

[0125] In the above formula, represents the signal variance, which is used to control the overall scaling of the function, and l represents the length scale parameter.

[0126] Specifically, at each time step x, the known time series SOH matrix is used as training data to train the Gaussian Process Regression model, and then the current time step x t is predicted to obtain μ t and wherein μ t is the prediction value of the time step, is the uncertainty estimate of the prediction result, and a smaller uncertainty estimate indicates that the model has a higher confidence in the current prediction, and vice versa.

[0127] S520, filtering the global sequence points of the health degree sequence after data enhancement by using the Gaussian Process Regression model to obtain the sequence estimate value.

[0128] In the embodiment of the present application, a double-layer Gaussian Process Regression model is used for filtering, and the trained Gaussian Process Regression model is used to predict the global sequence points in the health degree sequence after data enhancement to obtain the sequence estimate value r1.

[0129] S530, removing the noise points in the health degree sequence according to the sequence estimate value to obtain a non-noise point sequence.

[0130] Since there are noise points in the data-enhanced health degree sequence, when the data-enhanced health degree sequence satisfies |r i -r1|≥r0, the noise point is actively removed, wherein r i r1 represents a single-point health degree of the noise point, r1 represents a sequence estimation value, and r0 represents an estimation value threshold, which is set according to actual conditions. After repeating the entire sequence, a non-noise point sequence can be obtained.

[0131] S540, locally filtering the non-noise point sequence by using a Gaussian process regression model to obtain an optimal estimated health degree sequence.

[0132] The non-noise point sequence is locally filtered by using a second layer of the Gaussian process regression model to obtain an optimal estimated health degree sequence excluding the noise point.

[0133] Figure 8 is a result diagram of data filtering in the embodiment of the application, as shown in Figure 8 the data-enhanced health degree sequence is composed of a single-point health degree and a data-enhanced point, and the result of the double-layer filtering is an optimal estimated health degree sequence curve, from Figure 8 which it can be seen that the filtering mode has high confidence and is not affected by local noise points, and the prediction stability and precision reach a good level, and the generalization performance is greatly improved. At the same time, the decay trend of SOH is different under different working conditions, and the decay trend is not linearly decreasing with the cycle or calendar, but has a strong nonlinear relationship. This also shows that the filtering mode has good precision and stability, and as the battery is continuously used, the method can be applied to SOH prediction in the whole life cycle, and can be automatically updated and iterated.

[0134] S600, diagnosing the health state of the battery according to the optimal estimated health degree sequence.

[0135] In the embodiment of the application, the optimal estimated health degree sequences of all battery cells are combined to form a battery cell health degree list soh_list, and the SOH decay trend or outlier characteristics of each battery cell are tracked, so that the health state of the battery is diagnosed, and the battery cell with abnormal decay is warned. Figure 9 is a fifth schematic diagram of the battery health state diagnosis method provided by the embodiment of the application, as shown in Figure 9 the prediction of the health state of the battery includes the following steps:

[0136] S610, calculating the decay rate of each battery cell according to the optimal estimated health degree sequence.

[0137] In the embodiment of the present application, soh_list_1 represents the battery cell health degree list of the first time period, soh_list_2 represents the battery cell health degree list of the second time period, and the calculation formula of the decay rate is as follows:

[0138] soh_rate_list = soh_list_1 - soh_list_2

[0139] In the above formula, soh_rate_list represents the monthly decay rate list of each battery cell.

[0140] S620, calculate the health degree of each battery cell according to the optimal estimation health degree sequence.

[0141] In the embodiment of the present application, the health degree of each battery cell is calculated according to the battery cell health degree list soh_list, and the calculation formula is as follows:

[0142] soh_list_disp = (soh_list - soh_list_avg) / soh_list_std

[0143] In the above formula, soh_list_disp represents the health degree dispersion list, soh_list_avg represents the average value of the battery cell health degree list, and soh_list_std represents the standard deviation of the battery cell health degree list.

[0144] S630, diagnose the health state of the battery according to the decay rate and / or the health degree dispersion.

[0145] Optionally, if the decay rate soh_rate_list_i of the i-th battery cell exceeds the decay rate threshold, and / or if soh_list_max - soh_list_min exceeds the dispersion threshold and the health degree dispersion soh_list_disp_i of the i-th battery cell exceeds the dispersion threshold, the health state of the battery is alarmed, and the alarm information is pushed to the client, the specific abnormal battery cell is located and identified, and offline maintenance or update is performed.

[0146] As a preferred embodiment of the present application, Figure 10 is the sixth schematic diagram of the battery health state diagnosis method provided by the embodiment of the present application, as shown in the figure, the method further comprises the following steps: Figure 10

[0147] S700, according to the optimal estimation health degree sequence, the state of charge-open circuit voltage curve is parameter reconstructed to obtain the state of charge-open circuit voltage curve under the optimal health degree.

[0148] ​With the aging of the battery, the SOC-OCV curve will be significantly different, the initial SOC-OCV curve obtained in the laboratory cannot meet the needs of SOH estimation at all, therefore, the SOC-OCV curve of all battery cells needs to be reconstructed and verified to meet the SOH estimation in the whole life cycle. Figure 11 is the seventh schematic diagram of the battery health state diagnosis method provided by the embodiment of the present application, as shown in the figure, the parameter reconstruction specifically includes the following steps: Figure 11

[0149] S710, calibrate the state of charge of each battery cell according to the optimal estimated health degree sequence.

[0150] Since there is a large error between the SOC sent by the BMS and the real SOC, it is necessary to calibrate the SOC based on the optimal estimated health degree sequence according to the ampere-hour integral under the charge and discharge condition. Specifically, the charge state of discharge is calibrated according to the following formula:

[0151] SOC F = 1 - C F / (SOH i × C N )

[0152] In the above formula, SOC F represents the charge state of discharge, C F represents the discharge capacity, which is obtained by ampere-hour integration according to the current, SOH i represents the optimal estimated health degree of the i-th battery cell, and C N represents the rated capacity calibrated when the battery is shipped.

[0153] The charge state of charge is calibrated according to the following formula:

[0154] SOC C = SOC0 + C C / (SOH i × C N )

[0155] In the above formula, SOC C represents the charge state of charge, SOC0 represents the initial value of the charge state of discharge at the last time, C C represents the charge capacity, which is obtained by ampere-hour integration according to the current.

[0156] S720, according to the calibrated state of charge and the state of charge-open circuit voltage curve, the open circuit voltage value of each battery cell is calculated to obtain the state of charge-open circuit voltage matrix.

[0157] ​According to the state of charge of all the calibrated electric cores, the open circuit voltage value of each electric core is counted through the state of charge-open circuit voltage curve, and a state of charge-open circuit voltage matrix in a big data scene is obtained.

[0158] In S730, the state of charge-open circuit voltage matrix is aggregated to obtain a state of charge-open circuit voltage curve in an optimal health state.

[0159] The state of charge-open circuit voltage matrix is aggregated in big data, the state of charge-open circuit voltage curve is reacquired and corrected, and a state of charge-open circuit voltage curve in an optimal health state is obtained. Figure 12 is a schematic diagram of the state of charge-open circuit voltage curve reconstruction in the embodiment of the application, as Figure 12 shown, the closed-loop update iteration is performed every 5% attenuation, the curve obviously deviates as the SOH decreases from high to low, which will obviously affect the estimation of the SOH, and the initial curve cannot meet the subsequent SOH estimation at all.

[0160] The battery health state diagnosis method provided in the embodiment is verified on a battery sample, and the root mean square error RMSE of a large number of battery samples under real working conditions is ≤2%, which is improved by more than 30% compared with the existing method, and the method can also generalize different types of batteries. The method uses the estimated optimal health state to complete the dynamic reconstruction of the state of charge-open circuit voltage curve, realizes the self-update of the model parameters, and through the data enhancement method for sparse data under real working conditions, the model generalization effect and stability under real working conditions are significantly improved. Based on the double-layer filtering mode, the adaptive update mode of the health state is realized, and the health state estimation in the whole life cycle under real working conditions is generalized.

[0161] Embodiment 2

[0162] Based on the same technical concept, the embodiment of the application provides a battery health state diagnosis device, Figure 13 is a result schematic diagram of the battery health state diagnosis device provided by the embodiment of the application, as Figure 13 shown, the battery health state diagnosis device 800 comprises:

[0163] The data acquisition module 810 is configured to acquire battery data for data preprocessing to obtain a standard data set.

[0164] The curve construction module 820 is configured to construct a state of charge-open circuit voltage curve of the battery core.

[0165] The data estimation module 830 is configured to estimate the single-point health degree of each electric core in a single charging segment according to the standard data set and the state of charge-open circuit voltage curve.

[0166] The data enhancement module 840 is configured to obtain a health degree sequence according to the single-point health degree, and perform data enhancement on the health degree sequence.

[0167] The data filtering module 850 is configured to perform double-layer filtering on the health degree sequence after data enhancement to obtain an optimal estimated health degree sequence.

[0168] The state diagnosis module 860 is configured to diagnose the health state of the battery according to the optimal estimated health degree sequence.

[0169] The battery health state diagnosis device provided by the embodiment of the present application can generalize the health stable estimation of the whole life cycle under real working conditions, and can generalize the batteries of different types of batteries, thereby effectively improving the accuracy, stability and generalization of the model. The health diagnosis of the battery cells or modules in the battery pack can be realized, and the health state diagnosis at the cell level also provides strong support for the maintenance and abnormal diagnosis of the battery cells or modules.

[0170] It can be understood that the implementation manners of the battery health state diagnosis method described in Embodiment 1 are also applicable to the present embodiment and can achieve the same technical effects, and therefore will not be described here.

[0171] Embodiment 3

[0172] Based on the same idea, the embodiment of the present application also provides an electronic device, Figure 14 is a structural schematic diagram of an electronic device provided by the embodiment of the present application, as Figure 14 indicated, the electronic device 900 can include a processor 910, a communications interface 920, a memory 930 and a communications bus 940, wherein the processor 910, the communications interface 920 and the memory 930 complete mutual communication through the communications bus 940. The processor 910 can invoke the logical instructions in the memory 930 to execute the steps of the battery health state diagnosis method described in the above embodiments. For example, it includes:

[0173] S100, obtaining battery data for data preprocessing to obtain a standard data set;

[0174] S200, constructing a state of charge-open circuit voltage curve of the battery cell;

[0175] S300, estimating the single-point health degree of each battery cell in a single charging segment according to the standard data set and the state of charge-open circuit voltage curve;

[0176] S400, obtaining a health degree sequence according to the single-point health degree, and performing data enhancement on the health degree sequence;

[0177] S500, performing double-layer filtering on the health degree sequence after data enhancement to obtain an optimal estimated health degree sequence;

[0178] S600, diagnosing the health state of the battery according to the optimal estimated health degree sequence.

[0179] The processor 910 can be a central processing unit (CPU). The processor can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, or a combination of the above.

[0180] In addition, the logic instructions in the memory 930 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0181] The memory 930 can include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function; the data storage area can store data created by the processor, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged with respect to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0182] Example 4

[0183] Based on the same concept, the embodiments of the present application also provide a computer readable storage medium, which stores a computer program containing at least one code, the at least one code being executable by a host device to control the host device to implement the steps of the battery health state diagnosis method as described in the above embodiments. For example, it includes:

[0184] S100, obtaining battery data for data preprocessing to obtain a standard data set;

[0185] S200, constructing a state of charge-open circuit voltage curve of the battery cell;

[0186] S300, estimating the single-point health degree of each cell in a single charging segment according to the standard data set and the state of charge-open circuit voltage curve;

[0187] S400, obtaining a health degree sequence according to the single-point health degree, and performing data enhancement on the health degree sequence;

[0188] S500, performing double-layer filtering on the health degree sequence after data enhancement to obtain an optimal estimated health degree sequence;

[0189] S600, diagnosing the health state of the battery according to the optimal estimated health degree sequence.

[0190] Based on the same technical concept, the embodiments of the present application also provide a computer program, which, when executed by a host device, is used to implement the above method embodiments.

[0191] The computer program can be stored in whole or in part on a computer readable storage medium packaged together with the processor, or in part or in whole on a memory not packaged together with the processor.

[0192] Based on the same technical concept, the embodiments of the present application also provide a processor, which is used to implement the above method embodiments. The above processor can be a chip.

[0193] In summary, the battery health state diagnosis method, device, electronic equipment and storage medium provided by the present application adopt the method of dynamically reconstructing the state of charge-open circuit voltage curve by estimating the optimal health state degree, realize the self-update of model parameters, and through the data enhancement method for sparse data under real working conditions, significantly improve the model generalization effect and stability under real working conditions. Based on the double-layer filtering mode, the adaptive updating mode of the health state degree is realized, which can generalize the health stable estimation of the whole life cycle under real working conditions, and can also generalize the batteries of different battery types, effectively improving the accuracy, stability and generalization of the model. The health diagnosis of the cells or modules in the battery pack can be realized, and the cell-level health state diagnosis also provides strong support for the maintenance and abnormal diagnosis of the cells or modules.

[0194] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. As will be apparent to those of ordinary skill in the art, embodiments described herein can be combined with other embodiments in accordance with the application.

[0195] The above embodiments only express several implementation ways of the present application, which are described in detail and specifically, but can not be understood as the limitation to the patent scope of the present application. It should be pointed out that, for the ordinary skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, which are all within the protection scope of the present application. Therefore, the patent protection scope of the present application should be subject to the appended claims.

[0196] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for diagnosing battery health status, characterized in that, The method includes: Obtain battery data and perform data preprocessing to obtain a standard dataset; Construct the state-of-charge-open-circuit voltage curve of the battery cell; The single-point health of each cell in a single charging segment is estimated based on the standard dataset and the state-of-charge-open-circuit voltage curve. A health score sequence is obtained based on the single-point health score, and data augmentation is performed on the health score sequence; The data-augmented health sequence is subjected to a two-layer filter to obtain the optimal estimated health sequence. Based on the optimal estimated health sequence, the state-of-charge-open-circuit voltage curve is reconstructed to obtain the state-of-charge-open-circuit voltage curve under the optimal health state, including: The state of charge of each cell is calibrated based on the optimal estimated health sequence and the ampere-hour integral under charge and discharge conditions. Based on the calibrated state of charge and the state of charge-open circuit voltage curve, the open circuit voltage value of each cell is calculated to obtain the state of charge-open circuit voltage matrix; The state-of-charge-open-circuit voltage matrix is ​​aggregated to obtain the state-of-charge-open-circuit voltage curve under the optimal health state; The health status of the battery is diagnosed based on the optimal estimated health sequence.

2. The battery health status diagnosis method according to claim 1, characterized in that, The step of estimating the single-point health of each cell in a single charging segment based on the state-of-charge-open-circuit voltage curve includes: Acquire charging segments that include a stationary state at the beginning and end of the charging process; The first voltage of each cell at the start of charging under the charging segment is input into the state of charge-open circuit voltage curve to obtain the first state of charge of each cell. The second voltage of each cell at the end of the charging segment is input into the state of charge-open circuit voltage curve to obtain the second state of charge of each cell. The real-time charging capacity of the charging segment is obtained by accumulating the current in ampere-hours and integrating it. The single-point health of each cell is calculated based on the first state of charge, the second state of charge, and the real-time charging capacity.

3. The battery health status diagnosis method according to claim 2, characterized in that, The step of obtaining a health score sequence based on the single-point health score and performing data augmentation on the health score sequence includes: Construct a variational autoencoder model; A time-series health matrix is ​​obtained by performing time-series calculations based on the single-point health scores. The variational autoencoder model is trained based on the time-series health matrix to obtain the optimal variational autoencoder model. The health score sequence is constructed based on the single-point health score, and the health score sequence is input into the optimal variational autoencoder model for data augmentation.

4. The battery health status diagnosis method according to claim 3, characterized in that, The step of performing a two-layer filter on the data-enhanced health sequence to obtain the optimal estimated health sequence includes: Construct a Gaussian process regression model; The global sequence points of the data-augmented health sequence are filtered using the Gaussian process regression model to obtain the sequence estimate. Based on the sequence estimate, noise points in the health sequence are removed to obtain a noise-free sequence. The Gaussian process regression model is used to perform local filtering on the non-noise point sequence to obtain the optimal estimated health sequence.

5. The battery health status diagnosis method according to claim 1, characterized in that, The step of diagnosing the battery's health status based on the optimal estimated health sequence includes: The attenuation rate of each cell is calculated based on the optimal estimated health sequence; Calculate the health outlier of each cell based on the optimal estimated health sequence; The health status of the battery is diagnosed based on the degradation rate and / or the health outlier.

6. A battery health status diagnostic device, characterized in that, The device includes: The data acquisition module is used to acquire battery data, perform data preprocessing, and obtain a standard dataset. The curve construction module is used to construct the state-of-charge-open-circuit voltage curve of a battery cell. The data estimation module is used to estimate the single-point health of each cell in a single charging segment based on the standard dataset and the state-of-charge-open-circuit voltage curve. The data augmentation module is used to obtain a health sequence based on the single-point health score and to augment the health sequence. The data filtering module is used to perform two-layer filtering on the data-enhanced health sequence to obtain the optimal estimated health sequence. The status diagnosis module is used to diagnose the health status of the battery based on the optimal estimated health sequence. The curve construction module is further configured to reconstruct the state-of-charge-open-circuit voltage curve based on the optimal estimated health sequence, to obtain the state-of-charge-open-circuit voltage curve under the optimal health state, including: The state of charge of each cell is calibrated based on the optimal estimated health sequence and the ampere-hour integral under charge and discharge conditions. Based on the calibrated state of charge and the state of charge-open circuit voltage curve, the open circuit voltage value of each cell is calculated to obtain the state of charge-open circuit voltage matrix; The state-of-charge-open-circuit voltage matrix is ​​aggregated to obtain the state-of-charge-open-circuit voltage curve under the optimal health state.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the battery health status diagnosis method as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the battery health status diagnosis method as described in any one of claims 1 to 5.

Images