An online estimation method for maximum battery capacity based on partial discharge data

Abstract

The present invention relates to an online estimation method for the maximum capacity of a battery based on partial discharge data, and belongs to the field of battery health management. The method comprises the following steps: S1: selecting a battery to be tested, obtaining the basic parameters of the battery and calibrating the initial capacity; S2: performing a cyclic charge and discharge experiment on the battery with an aging protocol of constant current and constant voltage charging and constant current and discharging, collecting the current and voltage data during the experiment, and preprocessing the data to establish a battery aging data set. S3: calculating the discharge capacity of each cycle, extracting the capacity corresponding to a specific voltage during the battery discharge process as a health factor, and constructing a battery capacity estimation model in combination with a Gaussian process regression algorithm. S4: using a trained Gaussian process regression model to perform online estimation of the battery capacity. The present invention utilizes partial discharge capacity data to avoid problems that may occur in the process of extracting health factors based on IC curves.

Description

Technical Field

[0001] The invention belongs to the field of battery health management and relates to an online estimation method for maximum battery capacity based on partial discharge data. Background Art

[0002] Although new energy technologies have made great progress in the past few years, more powerful measures are still needed to promote the transformation of clean energy. Lithium-ion batteries have been widely used in both mobile and fixed energy storage applications due to their many advantages. However, like other electrochemical systems, lithium-ion batteries will inevitably age with continuous use. When they age to a certain extent, the safety performance of the battery will be difficult to guarantee. In severe cases, they may even cause economic losses and casualties. Therefore, it is necessary to evaluate their health status. The incremental capacity (IC) curve method is one of the most advanced battery capacity evaluation methods. The large number of features contained in it, also known as health indicators (HI), have been proven to have a strong correlation with battery capacity, but the identification of most features cannot be separated from the IC curve peak. However, there may be many problems in this process. First, the calculation process of the IC curve requires differential operation, which makes it extremely sensitive to measurement noise, and the selection of different voltage intervals will also lead to very different shapes of the IC curve. In addition, in order to facilitate the tracking of features in the IC curve, smoothing the IC curve is also an indispensable step. Unfortunately, using the same smoothing window parameters throughout the battery life cycle may result in the IC peak not appearing in the later stages of battery aging. Therefore, choosing a reasonable voltage interval and smoothing window is a prominent difficulty in the IC curve capacity estimation process, and there is no good solution to the above problem in existing research. Summary of the invention

[0003] In view of this, an object of the present invention is to provide an online estimation method for the maximum capacity of a battery based on partial discharge data.

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

[0005] An online estimation method for the maximum capacity of a battery based on partial discharge data, the method extracts health factors from the partial discharge capacity data, uses a Gaussian process regression model to construct a mapping relationship between the health factor and the maximum capacity, and uses a trained Gaussian process regression model to achieve accurate estimation of the battery cell capacity, comprising the following steps:

[0006] S1: Select the battery to be tested, obtain the basic parameters of the battery and calibrate the initial capacity;

[0007] S2: Perform a cyclic charge-discharge experiment on the battery with a constant current and constant voltage charging and constant current and discharging protocol, collect the current and voltage data during the experiment, and pre-process the data to establish a battery aging data set;

[0008] S3: Extract the capacity corresponding to a specific voltage during battery discharge as a health factor, and build a battery capacity estimation model using a Gaussian process regression algorithm;

[0009] S4: Online estimation of battery capacity using the trained Gaussian process regression model.

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

[0011] S11: Select the battery to be tested and obtain the basic parameters of the battery;

[0012] S12: Place the selected battery in a constant temperature box at 25°C for 2 hours to allow it to reach thermal equilibrium;

[0013] S13: The battery is charged at a constant current rate of 0.2C (charge and discharge current / nominal capacity). When the battery voltage reaches the upper cut-off voltage of 3.65V, it is changed to constant voltage charging until the charging current is less than or equal to 0.05C. Charging is stopped. At this time, the battery is considered to be fully charged.

[0014] S14: The battery is left to stand for 30 minutes to allow the battery to reach thermal equilibrium;

[0015] S15: The battery is discharged at a constant current of 0.2C to a lower cut-off voltage of 2.5V. At this point, the battery is considered fully discharged, and the amount of electricity released by the battery in this process is recorded;

[0016] S16: The battery is left to stand for 30 minutes to allow the battery to reach thermal equilibrium;

[0017] S17: Repeat S13 to S16 three times, and calculate the average value of the three measured discharge capacities as the final initial capacity of the battery.

[0018] Further, the S2 specifically includes the following steps:

[0019] S21: Place the battery in a 35°C constant temperature box for 2 hours to allow it to reach thermal equilibrium;

[0020] S22: The battery is charged at a constant current rate of 0.3C. When the battery voltage reaches the upper cut-off voltage of 3.65V, it is changed to constant voltage charging until the charging current is less than or equal to 0.05C. Charging is stopped. At this time, the battery is considered to be fully charged.

[0021] S23: The battery is left to stand for 10 minutes;

[0022] S24: The battery is discharged at a constant current of 1C to a lower cut-off voltage of 2.5V, at which point the battery is considered fully discharged;

[0023] S25: The battery is left to stand for 10 minutes;

[0024] S26: Repeat S22 to S25 until the battery capacity decays to 80% of its initial value;

[0025] S27: Preprocess the current and voltage data recorded during the experiment, including unifying units, cycle slicing, eliminating incomplete cycles, and establishing a battery cycle aging data set.

[0026] Furthermore, the constant current and constant voltage charging and discharging conditions set in the battery aging experiment in S2 are variable, including different charging and discharging rates and aging temperatures.

[0027] Further, the S3 specifically includes the following steps:

[0028] S31: Integrate the current-time data recorded in the experiment in ampere-hours to calculate the maximum capacity of each cycle of the battery;

[0029] S32: Calculate the battery capacity corresponding to each voltage in all discharge cycles in the same way to obtain a discharge capacity curve in each cycle, i.e., a discharge capacity-voltage curve;

[0030] S33: extracting the capacity corresponding to a specific voltage from the discharge capacity curve as a health factor (the health factor avoids a series of preprocessing processes, including differentiation, smoothing and other operations, and the health factor can be approximately equivalent to the area under the IC curve);

[0031] S34: Evaluate the health factor, and calculate the Pearson correlation coefficient between the maximum capacity and each dimension of the health factor to determine the voltage extraction interval of the health factor;

[0032] S35: The extracted health factor and the maximum capacity of the battery are used as input and output of the Gaussian process regression algorithm to train it.

[0033] Furthermore, the health factor evaluation method selected for the battery in S34 is the Pearson correlation coefficient method, but is not limited to this evaluation standard, and may also be the Spearman coefficient method, the grey correlation method, etc.; the battery capacity estimation model selected in S35 is the Gaussian process regression algorithm, but is not limited to this machine learning method, and may also be multi-layer perceptron, random forest, support vector machine and other algorithms.

[0034] Furthermore, S4 is specifically as follows: extracting the same health factors in the test set batteries as those in the training set batteries, and inputting them into the trained Gaussian process regression model, so as to estimate the maximum capacity of the batteries in the test set.

[0035] Furthermore, the health factor avoids the problems that may occur in the process of extracting the health factor based on the IC curve, and is expected to be able to train a GPR model with better estimation effect on less training set data.

[0036] The beneficial effects of the present invention are:

[0037] (1) Extracting health factors from discharge conditions is more in line with the actual needs of energy storage batteries.

[0038] (2) The health factors proposed in the present invention can be extracted using raw data without the need for complex preprocessing, thereby reducing the burden of calculation or storage.

[0039] (3) The proposed health factor is homologous to the health factor in the IC curve, but avoids the selection of subjective parameters in the IC curve calculation process, such as voltage interval and smoothing window, and can be extracted from part of the discharge process.

[0040] (4) The health factor proposed in the present invention has a strong correlation with the battery capacity. The capacity estimation model trained using the early aging data of the battery may have a better capacity estimation performance.

[0041] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0043] Figure 1 It is the overall logic diagram of the method of the present invention;

[0044] Figure 2 A schematic diagram of the extraction process of the health factor in the invention;

[0045] Figure 3 The figure is a scatter plot of the health factor and the maximum capacity of the battery proposed in the present invention and the correlation coefficient. DETAILED DESCRIPTION

[0046] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0047] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0048] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0049] See also Figure 1 ,The online estimation method of the maximum capacity of the battery based on partial discharge data can be divided into the following steps:

[0050] S1: Select the battery to be tested, obtain the basic parameters of the battery and calibrate the initial capacity;

[0051] S2: Perform a cyclic charge-discharge experiment on the battery with a constant current and constant voltage charging and constant current and discharging protocol, collect the current and voltage data during the experiment, and pre-process the data to establish a battery aging data set;

[0052] S3: Extract the capacity corresponding to a specific voltage during battery discharge as a health factor, and build a battery capacity estimation model using a Gaussian process regression algorithm;

[0053] S4: Online estimation of battery capacity using the trained Gaussian process regression model.

[0054] As an optional embodiment, the above S1 specifically includes S11-S17:

[0055] S11: Select the battery to be tested and obtain the basic parameters of the battery;

[0056] S12: Place the selected battery in a constant temperature box at 25°C for 2 hours to allow it to reach thermal equilibrium;

[0057] S13: The battery is charged at a constant current rate of 0.2C (charge and discharge current / nominal capacity). When the battery voltage reaches the upper cut-off voltage of 3.65V, it is changed to constant voltage charging until the charging current is less than or equal to 0.05C. Charging is stopped. At this time, the battery is considered to be fully charged.

[0058] S14: The battery is left to stand for 30 minutes to allow the battery to reach thermal equilibrium;

[0059] S15: The battery is discharged at a constant current of 0.2C to a lower cut-off voltage of 2.5V. At this point, the battery is considered fully discharged, and the amount of electricity released by the battery in this process is recorded;

[0060] S16: The battery is left to stand for 30 minutes to allow the battery to reach thermal equilibrium;

[0061] S17: Repeat S13 to S16 three times, and calculate the average value of the three measured discharge capacities as the final initial capacity of the battery.

[0062] As an optional embodiment, the above S2 specifically includes S21-S27:

[0063] S21: Place the battery in a 35°C constant temperature box for 2 hours to allow it to reach thermal equilibrium;

[0064] S22: The battery is charged at a constant current rate of 0.3C. When the battery voltage reaches the upper cut-off voltage of 3.65V, it is changed to constant voltage charging until the charging current is less than or equal to 0.05C. Charging is stopped. At this time, the battery is considered to be fully charged.

[0065] S23: The battery is left to stand for 10 minutes;

[0066] S24: The battery is discharged at a constant current of 1C to a lower cut-off voltage of 2.5V, at which point the battery is considered fully discharged;

[0067] S25: The battery is left to stand for 10 minutes;

[0068] S26: Repeat S22 to S25 until the battery capacity decays to 80% of its initial value;

[0069] S27: Preprocess the current and voltage data recorded during the experiment, including unifying units, cycle slicing, eliminating incomplete cycles, and establishing a battery cycle aging data set.

[0070] The constant current and constant voltage charging and discharging conditions set in the battery aging experiment in the above steps are variable, including different charging and discharging rates and aging temperatures.

[0071] As an optional embodiment, the above S3 specifically includes S31-S35:

[0072] S31: Integrate the current-time data recorded in the experiment in ampere-hours, and the expression is:

[0073]

[0074] Where I is the discharge current, t is the battery discharge time, the maximum capacity vector of the battery is calculated, and the maximum capacity of the i-th cycle is recorded as Y i ;

[0075] S32: The battery capacity corresponding to each voltage in all discharge cycles is calculated in the same way to obtain the discharge capacity vector [V, Q(V)] in each cycle, and its specific expression is as follows:

[0076] V=[V start ,V start -ΔV,...,V end ]

[0077] Q=[Q 1 ,Q 2 ,...,Q k ] T

[0078] N=(V start -V end ) / ΔV+1

[0079] Where ΔV is the voltage sampling interval, k is the total number of cycles, and the starting voltage (V start ) and the termination voltage (V end ) is evenly divided into N-1 parts, and the capacity sequence of the i-th cycle is expressed as It is worth noting that the voltage sequence of all cycles remains the same, [V start ,V start +ΔV,...,V end ] However, since the data measured during the discharge process may not contain all the voltages in the voltage sequence, it is necessary to interpolate the original QV sequence. The method used can be linear interpolation or other interpolation methods.

[0080] S33: Figure 2The schematic diagram of the health factor extraction process is shown. The capacity corresponding to a specific voltage is extracted from the discharge capacity curve as the health factor (this health factor avoids a series of preprocessing processes, including differentiation, smoothing and other operations, and the health factor can be approximately equivalent to the area under the IC curve). The expression corresponding to the health factor of the ith cycle is X i ={x 0 ,x 1 ,…x m}, m is the index of voltage;

[0081] S34: Evaluate the health factor and calculate the Pearson correlation coefficient between the maximum capacity and each dimension of the health factor. The calculation formula is as follows:

[0082]

[0083] where p i is the health factor corresponding to the i-th cycle at a specific voltage, is the mean value of the health factor, q i is the i-th capacity, is the mean value of capacity, and k is the total number of battery cycles. The voltage extraction interval of the health factor is determined using the calculated Pearson correlation coefficient. It is recommended to select the voltage interval corresponding to the health factor with a Pearson correlation coefficient greater than 0.95. The evaluation index for the health factor can also be the Spearman coefficient, gray correlation degree, etc. Figure 3 The scatter plot of the extracted health factor and the maximum capacity of the battery at a specific voltage and the correlation coefficient between the two are shown;

[0084] S35: The extracted health factor and the maximum battery capacity are used as the input and output of the Gaussian process regression algorithm to train it. The input and output pairs used to train the model are {(X 1 ,Y 1 ),(X 2 ,Y 2 ),…,(X i ,Y i )}. As an optional special case, the kernel function in GPR is expressed as follows:

[0085] k(x,x')=x T x'+(x T x'+c) d

[0086] Where x,x'∈R dis any random variable, and the mean function is selected as 0. The above kernel function and mean function can be selected according to actual needs. The battery capacity estimation model can also be a multi-layer perceptron, random forest, support vector machine and other algorithms. The data used for training can be either the early aging data of the battery or the full life cycle data of the battery under the same working conditions;

[0087] As an optional embodiment, step S4 uses the GPR model trained in step S3 for capacity estimation. Specifically, in the capacity estimation stage, the battery needs to start constant current discharge from a fully charged state (in fact, this is possible because in most cases, the battery will be artificially charged), and the battery management system starts recording the voltage and calculating its corresponding capacity. When the battery voltage covers the X constructed during training, i The data can be stopped after the voltage range is reached. By inputting it into the trained model, the battery capacity under the cycle can be output. By further adjusting the voltage range, the capacity at the early stage of discharge under an unknown cycle can be used to estimate the capacity at the later stage of discharge.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. An online estimation method for the maximum capacity of a battery based on partial discharge data, characterized in that: The method comprises the following steps: S1: Select the battery to be tested, obtain the basic parameters of the battery and calibrate the initial capacity; S2: Perform a cyclic charge-discharge experiment on the battery with a constant current and constant voltage charging and constant current and discharging protocol, collect the current and voltage data during the experiment, and pre-process the data to establish a battery aging data set; S3: Extract the capacity corresponding to a specific voltage during battery discharge as a health factor, and build a battery capacity estimation model in combination with the Gaussian Process Regression (GPR) algorithm; this health factor avoids a series of preprocessing processes, including differentiation and smoothing operations, and is approximately equivalent to the area under the IC curve; S4: Use the trained Gaussian process regression model to estimate the battery capacity online; use partial discharge capacity data to avoid problems that may occur in the health factor extraction process based on IC curves; The S1 specifically includes the following steps: S11: Select the battery to be tested and obtain the basic parameters of the battery; S12: Place the selected battery in a constant temperature box at 25°C for 2 hours to allow it to reach thermal equilibrium; S13: The battery is charged at a constant current rate of 0.2C, where 0.2C is equal to the charge and discharge current divided by the nominal capacity. When the battery voltage reaches the upper cut-off voltage of 3.65V, it is changed to constant voltage charging until the charging current is less than or equal to 0.05C. Charging is stopped at this time. At this time, the battery is considered to be fully charged. S14: The battery is left to stand for 30 minutes to allow the battery to reach thermal equilibrium; S15: The battery is discharged at a constant current of 0.2C to a lower cut-off voltage of 2.5V. At this point, the battery is considered fully discharged, and the amount of electricity released by the battery in this process is recorded; S16: The battery is left to stand for 30 minutes to allow the battery to reach thermal equilibrium; S17: Repeat S13 to S16 three times, and calculate the average value of the three times of discharge capacity as the final initial capacity of the battery; The S2 specifically includes the following steps: S21: Place the battery in a 35°C constant temperature box for 2 hours to allow it to reach thermal equilibrium; S22: The battery is charged at a constant current rate of 0.3C. When the battery voltage reaches the upper cut-off voltage of 3.65V, it is changed to constant voltage charging until the charging current is less than or equal to 0.05C. Charging is stopped. At this time, the battery is considered to be fully charged. S23: The battery is left to stand for 10 minutes; S24: The battery is discharged at a constant current of 1C to a lower cut-off voltage of 2.5V, at which point the battery is considered fully discharged; S25: The battery is left to stand for 10 minutes; S26: Repeat S22 to S25 until the battery capacity decays to 80% of its initial value; S27: Preprocess the current and voltage data recorded during the experiment, including unifying units, cycle slicing, removing incomplete cycles, and establishing a battery cycle aging data set; In S2, the constant current and constant voltage charging and discharging conditions set in the battery aging experiment are variable, including different charging and discharging rates and aging temperatures; The S3 specifically includes the following steps: S31: Integrate the current-time data recorded in the experiment in ampere-hours to calculate the maximum capacity of each cycle of the battery; S32: Calculate the battery capacity corresponding to each voltage in all discharge cycles in the same way to obtain a discharge capacity curve in each cycle, i.e., a discharge capacity-voltage curve; S33: extracting the capacity corresponding to a specific voltage from the discharge capacity curve as a health factor; S34: Evaluate the health factor, and calculate the Pearson correlation coefficient between the maximum capacity and each dimension of the health factor to determine the voltage extraction interval of the health factor; S35: The extracted health factor and the maximum capacity of the battery are used as input and output of the Gaussian process regression algorithm to train it.

2. The method for online estimation of the maximum capacity of a battery based on partial discharge data according to claim 1, characterized in that: In S34, the health factor evaluation method selected for the battery is the Pearson correlation coefficient method, the Spearman coefficient method or the grey correlation method; in S35, the battery capacity estimation model selected is the Gaussian process regression algorithm, the multi-layer perceptron algorithm, the random forest algorithm or the support vector machine algorithm.

Images