Battery capacity estimation method, electronic device and storage medium

Abstract

The present application discloses a battery capacity estimation method including: obtaining the charging current, charging voltage, charging time, and charging voltage interval of the battery to be estimated; determining the charging current interval and charging time interval according to the charging current, the charging voltage, and the charging time, wherein the charging current interval and the charging time interval correspond to the charging voltage interval; obtaining training features when training the model, wherein the training features include training electrochemical features and training time domain features; determining the actual features of the battery to be estimated according to the charging voltage interval, the charging current interval, the charging time interval, and the training features, wherein the actual features include actual electrochemical features and actual time domain features; estimating the actual capacity of the battery to be estimated according to the model and the test features. The present application also provides an electronic device and a storage medium, which can reduce the demand for data volume and shorten the estimation time.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a battery capacity estimation method, an electronic device, and a storage medium. Background Art

[0002] In the early stage of using lithium-ion batteries, the accuracy of battery capacity estimation is relatively high. However, as the lithium-ion battery cells age, the accuracy of battery capacity estimation continues to decrease, which will cause the capacity displayed by electronic devices to jump, and even the electronic device may directly shut down when the displayed capacity is 20% to 30%. At present, the battery capacity can be estimated by using complete experimental cycle data. Although this method can improve the accuracy of battery capacity estimation, it requires a large amount of data and takes a long time. Summary of the invention

[0003] In view of this, it is necessary to provide a battery capacity estimation method, an electronic device and a storage medium, which can reduce the data volume requirement and shorten the estimation time.

[0004] An embodiment of the present application provides a battery capacity estimation method, the method comprising: obtaining a charging current, a charging voltage, a charging time, and a charging voltage range of a battery to be estimated; determining the charging current range and the charging time range according to the charging current, the charging voltage, and the charging time, wherein the charging current range and the charging time range correspond to the charging voltage range; obtaining training features when training a model, the training features comprising training electrochemical features and training time domain features; determining actual features of the battery to be estimated according to the charging voltage range, the charging current range, the charging time range, and the training features, the actual features comprising actual electrochemical features and actual time domain features; estimating the actual capacity of the battery to be estimated according to the model and the actual features.

[0005] According to some embodiments of the present application, the model is a Gaussian process regression model.

[0006] According to some embodiments of the present application, the difference between the maximum charging voltage and the minimum charging voltage in the charging voltage interval is less than or equal to 0.1 volt.

[0007] According to some embodiments of the present application, determining the actual characteristics of the battery to be estimated according to the charging voltage interval, the charging current interval, the charging time interval, and the training characteristics includes:

[0008] Determining the charging capacity of the battery to be estimated according to the charging current in the charging current interval and the charging time in the charging time interval;

[0009] Determining a capacity increment value according to the charging voltage in the charging voltage interval and the charged capacity;

[0010] The actual characteristics of the battery to be estimated are determined according to the charging current in the charging current interval, the charging voltage in the charging voltage interval, the charging time in the charging time interval, the capacity increment value, and the training characteristics.

[0011] According to some embodiments of the present application, before acquiring the training features when training the model, the method further includes:

[0012] Determine the total training characteristics under different cycle numbers according to the charging current, charging voltage and charging time of the standard battery under different cycle numbers and the charging voltage range, wherein the total training characteristics include the total training electrochemical characteristics and the total training time domain characteristics;

[0013] Determine a first total feature at different cycle numbers whose linear correlation between the total training feature at different cycle numbers and the actual capacity of the standard battery at different cycle numbers is greater than a preset value;

[0014] The training feature is determined according to the first total feature at different cycle times.

[0015] According to some embodiments of the present application, the total training electrochemical characteristics include activation energy increment, throughput and charging time, and the total training time domain characteristics include peak value, maximum value, mean, variance, root mean square value, peak factor, kurtosis, kurtosis factor, waveform factor, pulse factor, margin factor, minimum value and skewness factor.

[0016] According to some embodiments of the present application, before determining the training feature according to the first total feature at different numbers of cycles, the method further includes:

[0017] The total training features at different numbers of cycles are transformed into second total features at different numbers of cycles by using a principal component analysis method;

[0018] Determining the training feature according to the first total feature at different numbers of cycles includes:

[0019] Determine that the same feature in the first total feature at different cycle numbers and the second total feature at different cycle numbers is a third total feature at different cycle numbers;

[0020] The training feature used when determining the training model is the third total feature.

[0021] According to some embodiments of the present application, the method further includes:

[0022] Obtaining initial parameters generated when training the model using charging current, charging voltage, and charging time at different cycle times of a standard battery;

[0023] Acquire a first feature when the initial parameter is generated, the first feature comprising a first electrochemical feature and a first time domain feature;

[0024] Replacing the least relevant feature of the first feature with a second feature determined by the charging current, charging voltage and charging time at different cycle numbers in actual use of the battery to be estimated, wherein the second feature includes a second electrochemical feature and a second time domain feature;

[0025] The trained model is obtained by updating the initial parameters through the updated first feature.

[0026] According to some embodiments of the present application, the acquiring of the first feature when generating the initial parameter includes:

[0027] Acquire a first feature at a first number of cycles when generating the initial parameters;

[0028] The second feature determined by using the charging current, charging voltage and charging time at different cycle numbers in actual use of the battery to be estimated to replace the least relevant feature of the first feature includes:

[0029] Determine a second characteristic at a second number of cycles according to the first characteristic, the charging current, the charging voltage, and the charging time at different cycle numbers in actual use of the battery to be estimated;

[0030] Determine the smallest correlation at the second number of cycles among the correlations of the second feature at the second number of cycles and the first feature at the first number of cycles;

[0031] The feature corresponding to the correlation at the second number of cycles in the first feature at the first number of cycles is replaced by the second feature at the second number of cycles.

[0032] An embodiment of the present application provides an electronic device, the electronic device comprising:

[0033] Battery;

[0034] Processor; and

[0035] A memory, wherein a plurality of program modules are stored in the memory, and the plurality of program modules are loaded by the processor and execute the battery capacity estimation method as described in any one of the above items.

[0036] An embodiment of the present application provides a storage medium on which at least one computer instruction is stored, and the instruction is loaded by a processor to execute any of the battery capacity estimation methods described above.

[0037] The battery capacity estimation method, electronic device, and storage medium provided in the embodiments of the present application use less data to estimate the capacity of the battery to be estimated through a Gaussian process regression model based on the test electrochemical characteristics and test time domain characteristics generated according to the charging voltage range of the battery to be estimated, the charging current range corresponding to the charging voltage range, and the charging time range corresponding to the charging voltage range, thereby reducing the data volume requirement and shortening the estimation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a schematic diagram of an electronic device according to an embodiment of the present application.

[0039] Figure 2 This is a flow chart of a battery capacity estimation method according to an embodiment of the present application.

[0040] Figure 3 for Figure 2 Flowchart of determining training features for a battery capacity estimation method.

[0041] Figure 4 It is a linear correlation curve diagram of the total training characteristics at different cycle numbers and the actual capacity of the standard battery at different cycle numbers.

[0042] Figure 5 for Figure 2 The battery capacity estimation method updates the initial parameters to obtain a flow chart of the trained model.

[0043] Figure 6 for Figure 5 A flow chart of replacing the least relevant feature of the first feature with the first and second features determined by the charging current, charging voltage and charging time at different cycle numbers in actual use of the estimated battery.

[0044] Figure 7 for Figure 2 A flowchart for determining the actual characteristics of the battery to be estimated based on the charging voltage range, charging current range, charging time range, and training characteristics.

[0045] Main component symbols

[0046] Electronic device 100

[0047] Memory 11

[0048] Processor 12

[0049] Battery to be estimated 13

[0050] Analog to digital converter 14

[0051] Timer 15 DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0053] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] See also Figure 1 , Figure 1 The electronic device 100 includes, but is not limited to, a memory 11, at least one processor 12, a battery to be estimated 13, an analog-to-digital converter 14, and a timer 15, and the above components can be connected via a bus or directly.

[0055] It should be noted that Figure 1 The electronic device 100 is merely an example. In other embodiments, the electronic device 100 may include more or fewer components, or have a different component configuration. The electronic device 100 may be an electric motorcycle, an electric bicycle, an electric car, a mobile phone, a tablet computer, a digital assistant, a personal computer, or any other suitable rechargeable device.

[0056] In one embodiment, the battery to be estimated 13 is a rechargeable battery used to provide electrical energy to the electronic device 100. For example, the battery to be estimated 13 can be a lead-acid battery, a nickel-cadmium battery, a nickel-metal hydride battery, a polymer lithium-ion battery, a lithium polymer battery, and a lithium iron phosphate battery. The battery to be estimated 13 is logically connected to the processor 12 through a battery management system (BMS), so that the battery management system realizes functions such as charging and discharging. The battery management system can be connected to the energy storage inverter (PCS) via CAN or RS485. The battery to be estimated 13 includes a battery cell, and the battery to be estimated can be repeatedly charged in a cyclic and rechargeable manner.

[0057] In this embodiment, the analog-to-digital converter 14 is used to measure the charging voltage and charging current of the battery cell of the battery to be estimated 13 during the charging process. The timer 15 is used to record the charging time of the battery cell of the battery to be estimated 13 during the charging process. It is understandable that the electronic device 100 may also include other devices, such as a pressure sensor, a light sensor, a gyroscope, a hygrometer, an infrared sensor, etc.

[0058] See also Figure 2 , Figure 2 This is a flow chart of a battery capacity estimation method according to an embodiment of the present application. The battery capacity estimation method is applied to a battery to be estimated. The battery capacity estimation method comprises the following steps:

[0059] S21: Obtain the charging current, charging voltage, charging time, and charging voltage range of the battery to be estimated.

[0060] In this embodiment, the charging current and the charging voltage can be obtained through the analog-to-digital converter. The charging time can be obtained through the timer. Among them, the charging current, the charging voltage and the charging time are the charging current, charging voltage and charging time of the battery to be estimated in the constant current charging stage. The charging voltage interval is a value preset by the user, and the charging voltage interval is stored in the electronic device. Preferably, the charging voltage interval is [3.85, 3.95], and can also be other values, depending on specific needs. If the charging voltage interval is [3.85, 3.95], the maximum charging voltage is 3.95 volts, and the minimum charging voltage is 3.85 volts. In this embodiment, the difference between the maximum charging voltage and the minimum charging voltage in the charging voltage interval is less than or equal to 0.1 volt.

[0061] S22: determining a charging current interval and a charging time interval according to the charging current, the charging voltage and the charging time, wherein the charging current interval and the charging time interval correspond to the charging voltage interval.

[0062] In this embodiment, the voltage of the battery to be estimated continues to increase during the constant current charging stage, so that the charging current interval and the charging time interval are determined according to the charging current, the charging voltage and the charging time, and the charging current interval and the charging time interval data can be intercepted from the charging current, the charging voltage and the charging time data.

[0063] S23: Acquire training features when training the model, where the training features include training electrochemical features and training time domain features.

[0064] The model is a Gaussian process regression model. Figure 3As shown, before obtaining the training features when training the model, the battery capacity estimation method further includes:

[0065] S31: Determine the total training characteristics under different cycle numbers according to the charging current, charging voltage and charging time of the standard battery under different cycle numbers, and the charging voltage range, wherein the total training characteristics include the total training electrochemical characteristics and the total training time domain characteristics.

[0066] The method of determining the total training characteristics at different cycle times according to the charging current, charging voltage and charging time of the standard battery at different cycle times and the charging voltage interval includes:

[0067] b1: Obtain the charging current, charging voltage and charging time of the standard battery at different cycle times.

[0068] In this embodiment, each cycle number is a complete charge and discharge cycle. Preferably, the cycle number is in the interval [20, 200], and can also be other values, depending on specific needs.

[0069] The charging current, charging voltage and charging time of the standard battery at different cycle numbers may be stored in the electronic device or in an external device. Then, obtaining the charging current, charging voltage and charging time of the standard battery at different cycle numbers may be obtaining the charging current, charging voltage and charging time of the standard battery at different cycle numbers from the electronic device, or obtaining the charging current, charging voltage and charging time of the standard battery at different cycle numbers from an external device.

[0070] b2: Determine the charging current range and charging time range of the standard battery at different cycle numbers corresponding to the charging voltage range according to the charging current, charging voltage and charging time of the standard battery at different cycle numbers.

[0071] b3: Calculate the charging capacity of the standard battery according to the charging current within the charging current interval of the standard battery and the charging time within the charging time interval of the standard battery.

[0072] In this embodiment, the calculating the charging capacity of the standard battery according to the charging current within the charging current interval of the standard battery and the charging time within the charging time interval of the standard battery includes calculating the charging capacity of the standard battery by a Coulomb integration method.

[0073] b4: Calculate the capacity increment value according to the charging voltage within the charging voltage range of the standard battery and the charging capacity corresponding to the charging voltage.

[0074] In this embodiment, the capacity increment value is calculated based on the charging voltage within the charging voltage range of the standard battery and the charging capacity corresponding to the charging voltage, and the capacity increment value is calculated by differentiating the charging voltage within the charging voltage range of the standard battery and the charging capacity corresponding to the charging voltage.

[0075] b5: Draw a capacity increment curve of the standard battery at different cycle times according to the charging voltage within the charging voltage range of the standard battery and the capacity increment value.

[0076] In this embodiment, the capacity increment curve of the standard battery at different cycle times in the same coordinate system is plotted according to the charging voltage in the charging voltage range of the standard battery and the capacity increment value. Different capacity increment curves represent different cycle times. The abscissa of the capacity increment curve of the standard battery at different cycle times is the charging voltage, and the ordinate of the capacity increment curve of the standard battery at different cycle times is the capacity increment value.

[0077] b6: Determine the total training characteristics at different cycle numbers based on the capacity increment curve of the standard battery at different cycle numbers and the charging current interval and charging time interval of the standard battery at different cycle numbers.

[0078] In this embodiment, before determining the total training features at different numbers of cycles based on the capacity increment curve of the standard battery at different numbers of cycles and the charging current interval and charging time interval of the standard battery at different numbers of cycles, the method also includes: performing sliding filtering and Gaussian filtering on the capacity increment curve of the standard battery at different numbers of cycles.

[0079] The total training electrochemical characteristics include activation energy increment, throughput and charging time, and the total training time domain characteristics include peak value, maximum value, mean value, variance, root mean square value, peak factor, kurtosis, kurtosis factor, waveform factor, pulse factor, margin factor, minimum value and skewness factor.

[0080] In this embodiment, according to the formula Determine the activation energy increment at different cycle numbers. S is the activation energy increment at different cycle numbers, is the minimum charging time in the charging time interval under different cycle times of the standard battery, is the maximum charging time in the charging time interval under different cycle numbers of the standard battery, V is the charging voltage in the charging voltage interval under different cycle numbers of the standard battery, and t is the charging time in the charging time interval under different cycle numbers of the standard battery.

[0081] In this embodiment, according to the formula Determine the throughput at different cycle times. h is the throughput under different cycle numbers, is the minimum charging time in the charging time interval under different cycle times of the standard battery, is the maximum charging time in the charging time interval under different cycle numbers of the standard battery, I is the charging current in the charging current interval under different cycle numbers of the standard battery, and t is the charging time in the charging time interval under different cycle numbers of the standard battery.

[0082] In this embodiment, according to the formula Determine the charging time under different cycle numbers. Where ΔT is the charging time under different cycle numbers, is the maximum charging time in the charging time interval under different cycle times of the standard battery, It is the minimum charging time in the charging time intervals under different cycle numbers of the standard battery.

[0083] In this embodiment, the peak values ​​at different cycle times are determined according to the formula Peak=max(|dQ / dV|), wherein Peak is the peak value at different cycle times, and dQ / dV is the capacity increment value of the standard battery at different cycle times.

[0084] In this embodiment, the maximum value at different cycle times is determined according to the formula Max=max(dQ / dV), wherein Max is the maximum value at different cycle times, and dQ / dV is the capacity increment value of the standard battery at different cycle times.

[0085] In this embodiment, according to the formula Determine the mean under different cycle numbers. Wherein, Mean is the mean under different cycle numbers, n is the number of samplings of the charging voltage in the charging voltage range under different cycle numbers of the standard battery, i is the i-th sampling, dQ i / dV i is the i-th capacity increment value of the standard battery under different cycle numbers.

[0086] In this embodiment, according to the formula Determine the variance under different cycle numbers. Wherein, Var is the variance under different cycle numbers, n is the number of samplings of the charging voltage in the charging voltage interval under different cycle numbers of the standard battery, i is the i-th sampling, dQ i / dV i is the i-th capacity increment value of the standard battery under different cycle numbers, and Mean is the mean value under different cycle numbers.

[0087] In this embodiment, according to the formula Determine the root mean square value under different cycle numbers. Wherein, Rms is the root mean square value under different cycle numbers, n is the number of samplings of the charging voltage in the charging voltage range under different cycle numbers of the standard battery, i is the i-th sampling, dQ i / dV i is the i-th capacity increment value of the standard battery under different cycle numbers.

[0088] In this embodiment, according to the formula Determine the peak factor at different cycle numbers. Where CF is the peak factor at different cycle numbers, Peak is the maximum value at different cycle numbers, and Rms is the root mean square value at different cycle numbers.

[0089] In this embodiment, according to the formula Determine the kurtosis at different cycle numbers. Wherein, Kur is the kurtosis at different cycle numbers, n is the number of samplings of the charging voltage in the charging voltage interval at different cycle numbers of the standard battery, i is the i-th sampling, dQ i / dV i is the i-th capacity increment value of the standard battery under different cycle numbers.

[0090] In this embodiment, according to the formula Determine the kurtosis factor at different numbers of cycles. Where KF is the kurtosis factor at different numbers of cycles, Kur is the kurtosis at different numbers of cycles, and Rms is the root mean square value at different numbers of cycles.

[0091] In this embodiment, according to the formula Determine the waveform factor at different cycle numbers. Where FF is the waveform factor at different cycle numbers, Rms is the root mean square value at different cycle numbers, and Mean is the mean value at different cycle numbers.

[0092] In this embodiment, according to the formula Determine the pulse factor at different cycle numbers. Where IF is the pulse factor at different cycle numbers, Peak is the maximum value at different cycle numbers, and Mean is the mean value at different cycle numbers.

[0093] In this embodiment, according to the formula Determine the margin factor under different cycle numbers. Wherein, MF is the margin factor under different cycle numbers, Peak is the maximum value under different cycle numbers, n is the number of samplings of the charging voltage in the charging voltage range under different cycle numbers of the standard battery, i is the i-th sampling, dQ i / dV i is the i-th capacity increment value of the standard battery under different cycle numbers.

[0094] In this embodiment, the minimum value at different cycle times is determined according to the formula Min=min(|dQ / dV|), wherein Min is the minimum value at different cycle times, and dQ / dV is the capacity increment value of the standard battery at different cycle times.

[0095] In this embodiment, according to the formula Determine the skewness factor at different cycle numbers. Wherein SF is the skewness factor at different cycle numbers, n is the number of samplings of the charging voltage in the charging voltage range at different cycle numbers of the standard battery, i is the i-th sampling, dQ i / dV i is the i-th capacity increment value of the standard battery under different cycle numbers, and Mean is the mean value under different cycle numbers.

[0096] S32: Determine a first total feature at different cycle numbers whose linear correlation between the total training feature at different cycle numbers and the actual capacity of the standard battery at different cycle numbers is greater than a preset value.

[0097] In this embodiment, the actual capacity may be the actual capacity calculated by other systems in the electronic device, may be the actual capacity calculated by other devices outside the electronic device, or may be the actual capacity calculated by any known method used by this method.

[0098] In this embodiment, according to the formula To determine the linear correlation between the total training features at different cycle times and the actual capacity of the standard battery at different cycle times. k is the linear correlation between the value of the kth feature at each cycle number and the actual capacity at the cycle number, i is the i-th value of the kth feature at each cycle number or the total number of actual capacities, N is the value of the kth feature at each cycle number or the total number of actual capacities, X i is the value of the kth feature at the i-th cycle, is the average value of the kth feature under all cycles, Y i is the actual capacity of the kth feature at the i-th cycle number, is the average of the actual capacity of the kth feature under all cycle numbers.

[0099] Preferably, the preset value may be 0.8, or other values, depending on specific requirements.

[0100] like Figure 4 As shown, Figure 4The linear correlation of the activation energy increment, the throughput, the charging time, the peak value, the maximum value, the mean value, the variance, the root mean square value, the peak factor, the kurtosis, the kurtosis factor, the waveform factor, the pulse factor, the margin factor, the minimum value and the skewness factor is shown. Among them, the linear correlation of the activation energy increment, the throughput, the charging time, the peak value, the maximum value, the mean value, the variance, the root mean square value, the kurtosis and the minimum value is greater than 0.8. Then, the first total feature at different cycle times whose linear correlation of the total training features at different cycle times and the actual capacity of the standard battery at different cycle times is greater than the preset value includes: the features of the activation energy increment, the throughput, the charging time, the peak value, the maximum value, the mean value, the variance, the root mean square value, the kurtosis and the minimum value at different cycle times.

[0101] S33: Determine the training feature according to the first total feature at different numbers of cycles.

[0102] In this embodiment, before determining the training feature according to the first total feature under different cycle numbers, the battery capacity estimation method further includes:

[0103] The principal component analysis (PCA) method is used to transform the total training features under different cycle numbers into the second total features under different cycle numbers.

[0104] Determining the training feature according to the first total feature at different numbers of cycles includes:

[0105] c1: determining that the same feature among the first total feature under different cycle numbers and the second total feature under different cycle numbers is the third total feature under different cycle numbers;

[0106] c2: The training feature when determining the training model is the third total feature.

[0107] In this embodiment, the battery capacity estimation method further includes:

[0108] The model is trained using the training features.

[0109] In this embodiment, before training the model using the training features, the battery capacity estimation method further includes:

[0110] The training features are standardized to form first feature vectors under different numbers of cycles.

[0111] The training of the model by the training features comprises:

[0112] d1: randomly selecting first eigenvectors at a first preset number of cycles from the first eigenvectors at the cycles.

[0113] d2: The model is trained by the selected first feature vector to obtain initial parameters of the model, wherein the selected first feature vector is an input variable of the model, and the actual capacity corresponding to the selected first feature vector is an output variable of the model.

[0114] Preferably, the first preset number can be any value in the interval [15, 20], or other values, depending on specific needs. Preferably, the covariance function of the model is a linear kernel function K(x, y)=θ 2 x T y can also be other functions, depending on specific needs.

[0115] In this embodiment, the model, when given an input x, obtains the distribution of the function f(x), that is, y=f(x)+ε.

[0116] Among them, ε is the observed value disturbed by noise, to satisfy the mean of 0, the variance is The normal distribution of .

[0117] The set of observed values ​​of y obeys Gaussian distribution, that is, the prior distribution model:

[0118] Among them, Var(x) is the n×n covariance matrix, and the (i,j)th element is the linear kernel function, that is:

[0119] In this embodiment, the stochastic gradient descent method is used to iteratively update the parameter θ in the likelihood function to obtain the optimal parameter to train the model. The likelihood function is in the following form:

[0120]

[0121] Wherein, θ is the parameter of the model, x is the training data, Var(x) is the n×n covariance matrix, σ is the variance of the noise, I is the identity matrix, K is the kernel function, and n is the number of data in the training set. In this embodiment, if x is the selected first eigenvector, θ is the initial parameter of the model.

[0122] According to the theorem and the obtained prior distribution model, the corresponding posterior distribution model can be solved: f′~g(m(X),K(X));

[0123] Among them, X, x are different inputs, K is the kernel function, σ is the variance of the noise, I is the unit matrix, and y is the output value.

[0124] like Figure 5 As shown, in order to adjust the model parameters, the battery capacity estimation method further includes:

[0125] S51: Obtaining initial parameters generated when training the model using the charging current, charging voltage and charging time of a standard battery at different cycle times.

[0126] The initial parameters are initial parameters generated when the model is trained using the selected first feature vector.

[0127] S52: Acquire a first characteristic when the initial parameter is generated, wherein the first characteristic includes a first electrochemical characteristic and a first time domain characteristic.

[0128] S53: Replace the least relevant feature of the first features with a second feature determined by the charging current, charging voltage and charging time at different cycle numbers in actual use of the estimated battery, wherein the second feature includes a second electrochemical feature and a second time domain feature.

[0129] The obtaining of the first feature when generating the initial parameters includes: obtaining the first feature under the first number of cycles when generating the initial parameters. In this embodiment, the first number is the first preset number randomly selected. The first feature is the feature after the linear correlation is greater than the preset value and the principal component analysis. The type of the first feature is, for example, the activation energy increment, the charging time, the peak value, the maximum value, the mean, the variance, the root mean square value and the kurtosis.

[0130] like Figure 6 As shown, the second feature determined by using the charging current, charging voltage and charging time at different cycle numbers in actual use of the battery to be estimated to replace the least relevant feature of the first feature includes:

[0131] S61: Determine a second characteristic at a second number of cycles based on the first characteristic and the charging current, charging voltage and charging time at different cycle numbers in actual use of the estimated battery.

[0132] The second characteristic at the second number of cycles determined according to the first characteristic and the charging current, charging voltage and charging time at different cycle numbers in actual use of the estimated battery includes:

[0133] g1: Obtain the charging current, charging voltage and charging time at different cycle times in actual use of the estimated battery.

[0134] g2: Determine the charging current range and charging time range under different number of cycles in actual use of the battery to be estimated, corresponding to the charging voltage range, according to the charging current, charging voltage and charging time under different number of cycles in actual use of the battery to be estimated.

[0135] g3: Calculate the charging capacity of the battery to be estimated at different cycle numbers according to the charging current in the charging current interval under different cycle numbers of the battery to be estimated and the charging time in the charging time interval under different cycle numbers of the battery to be estimated.

[0136] In this embodiment, calculating the charging capacity of the battery to be estimated at different cycle numbers according to the charging current in the charging current range at different cycle numbers of the battery to be estimated and the charging time in the charging time range at different cycle numbers of the battery to be estimated includes calculating the charging capacity of the battery to be estimated at different cycle numbers by using the Coulomb integration method.

[0137] g4: Calculate the capacity increment value according to the charging voltage within the charging voltage range under different cycle numbers of the battery to be estimated and the charging capacity corresponding to the charging voltage.

[0138] In this embodiment, the capacity increment value is calculated based on the charging voltage in the charging voltage range under different cycle numbers of the battery to be estimated and the charged capacity corresponding to the charging voltage, and the capacity increment value is calculated by differentiating the charging voltage in the charging voltage range under different cycle numbers of the battery to be estimated and the charged capacity corresponding to the charging voltage.

[0139] g5: Draw a capacity increment curve for the battery to be estimated at different cycle numbers according to the charging voltage in the charging voltage range for the battery to be estimated at different cycle numbers and the capacity increment value.

[0140] In this embodiment, the capacity increment curve of the battery to be estimated at different cycle numbers is drawn in the same coordinate system according to the charging voltage in the charging voltage interval at different cycle numbers of the battery to be estimated and the capacity increment value. The abscissa of the capacity increment curve of the battery to be estimated at different cycle numbers is the charging voltage, and the ordinate of the capacity increment curve of the battery to be estimated at different cycle numbers is the capacity increment value.

[0141] g6: Determine the third characteristic at different cycle numbers based on the first characteristic, the capacity increment curve of the battery to be estimated at different cycle numbers, and the charging current range and charging time range of the battery to be estimated at different cycle numbers, wherein the third characteristic includes a third electrochemical characteristic and a third time domain characteristic.

[0142] In this embodiment, the type of the third feature is the same as the type of the first feature. For example, the first feature includes activation energy increment, charging time, peak value, maximum value, mean value, variance, root mean square value and kurtosis, and the activation energy increment, charging time, peak value, maximum value, mean value, variance, root mean square value and kurtosis under different cycle numbers are determined according to the capacity increment curve under different cycle numbers of the battery to be estimated and the charging current interval and charging time interval under different cycle numbers of the battery to be estimated.

[0143] g7: Randomly select the second features under a second preset number of cycles from the third features under the said number of cycles.

[0144] In this embodiment, the second preset number is smaller than the first preset number.

[0145] S62: Determine the smallest correlation at the second number of cycles among the correlations between the second feature at the second number of cycles and the first feature at the first number of cycles.

[0146] In this embodiment, before determining the correlation at the second number of cycles that is the smallest of the correlations of the second feature at the second number of cycles and the first feature at the first number of cycles, the battery capacity estimation method further includes:

[0147] The second features at the second number of cycles are normalized to form second feature vectors at the second number of cycles.

[0148] In this embodiment, determining the smallest correlation at the second number of cycles among the correlations of the second feature at the second number of cycles and the first feature at the first number of cycles includes:

[0149] h1: According to the formula Determine the correlation between the second eigenvector at each cycle number and the first eigenvector at each cycle number. ab is the linear correlation between the value of the second eigenvector at the ath cycle number and the value of the first eigenvector at the bth cycle number, i is the i-th value of the second eigenvector or the total number of values ​​of the first eigenvector at each cycle number, P is the total number of values ​​of the second eigenvector or the first eigenvector at each cycle number, P iis the value of the i-th second eigenvector at the a-th cycle, is the average value of the second eigenvector of the i-th loop under all the cycles, Q i is the value of the i-th first eigenvector at the b-th cycle, is the average value of the first eigenvector of the i-th loop under all the cycle numbers.

[0150] h2: Arrange the correlations into a matrix having a second number of rows and a first number of columns.

[0151] For example, if the first number is M and the second number is K, the correlations are arranged into a matrix

[0152] h3: Calculates the total value of each column in the matrix.

[0153] In this embodiment, r 11 …r 1M is the first column, r 1M …r KM The calculation of the total value of each column in the matrix may include, for example, calculating the total value r1 of the first column, calculating the total value r2 of the second column, ... calculating the total value r of the Kth column. K The calculation of the total value r1 of the first column may be, for example, by the formula r1=r 11 +r 12 +…+r 1M Calculate the total value r1 of the first column, where r1 is the total value of the first column, r 11 is the correlation of the first column and the first row, r 12 is the correlation of the first column and the second row, r 1M is the correlation of the first column and the Mth row.

[0154] h4: Determine the smallest second number of values ​​in the total value.

[0155] Determining the smallest second number of values ​​in the total value comprises:

[0156] I1: Sort the total values ​​according to a preset order, where the preset order includes from large to small or from small to large.

[0157] I2: Determine the smallest second number of values ​​in the total value according to the sorted total value.

[0158] For example, the total values ​​after sorting are 8, 7.5, 7.3, ..., 5.6, 5.3, 4.9, and the second quantity is 3, then the smallest second quantity values ​​are: 5.6, 5.3 and 4.9.

[0159] h5: Determine the correlation degree at the minimum second number of cycles according to the value of the minimum second number.

[0160] For example, the minimum second number of values ​​is the total value of the first column, the total value of the fifth column and the total value of the eighth column, then the correlation under the minimum second number of cycle numbers is the correlation under the first cycle number, the correlation under the fifth cycle number and the correlation under the eighth cycle number of the first number of cycle numbers.

[0161] S63: Replace the feature corresponding to the correlation at the second number of cycles in the first feature at the first number of cycles with the second feature at the second number of cycles.

[0162] S54: Update the initial parameters by using the updated first feature to obtain the trained model.

[0163] In this embodiment, the model is trained by training the model with the first feature under the updated first number of cycles to update the initial parameters to obtain the trained model, wherein the first feature under the updated first number of cycles is the input variable of the model, and the actual capacity corresponding to the first feature under the updated first number of cycles is the output variable of the model. S24: Determine the actual characteristics of the battery to be estimated according to the charging voltage interval, the charging current interval, the charging time interval, and the training characteristics.

[0164] like Figure 7 As shown, the determining the actual characteristics of the battery to be estimated according to the charging voltage interval, the charging current interval, the charging time interval, and the training characteristics includes:

[0165] S71: Determine the charging capacity of the battery to be estimated according to the charging current in the charging current interval and the charging time in the charging time interval.

[0166] In this embodiment, determining the charge capacity of the battery to be estimated according to the charging current in the charging current interval and the charging time in the charging time interval includes determining the charge capacity of the battery to be estimated by using a Coulomb integration method.

[0167] S72: Determine a capacity increment value according to the charging voltage in the charging voltage interval and the charged capacity.

[0168] In this embodiment, determining the capacity increment value according to the charging voltage in the charging voltage interval and the charged capacity is to determine the capacity increment value by differentiating the charging voltage in the charging voltage interval and the charged capacity corresponding to the charging voltage.

[0169] S73: Determine the actual characteristics of the battery to be estimated according to the charging current in the charging current interval, the charging voltage in the charging voltage interval, the charging time in the charging time interval, the capacity increment value, and the training characteristics.

[0170] In this embodiment, the type of the actual feature of the battery to be estimated is the same as the type of the training feature.

[0171] S25: estimating the actual capacity of the battery to be estimated based on the model and the actual characteristics.

[0172] In this embodiment, the actual features are input into the posterior distribution model of the trained model to output the predicted mean and predicted confidence interval of the actual capacity of the battery to be estimated. This case not only outputs the predicted mean of the actual capacity of the battery to be estimated, but also outputs the predicted confidence interval, so that the credibility of the predicted mean can be determined, providing more information for determining the actual capacity of the battery to be estimated.

[0173] In this case, the actual capacity of the electromagnetic to be estimated is estimated through the data volume, actual characteristics and model corresponding to the charging voltage interval. Only less data is needed to realize the estimation of the capacity of the battery to be estimated through the model, which reduces the demand for data volume and shortens the estimation time. At the same time, when training the model, only the training features generated by the data volume of the standard battery under different cycle times corresponding to the charging voltage interval are needed to train the model, which reduces the amount of training data and shortens the training time. In addition, the model can also be updated as needed during actual use, which can improve the estimation accuracy of the capacity of the battery to be estimated.

[0174] Please continue reading Figure 1 In this embodiment, the memory 11 may be an internal memory of the electronic device, that is, a memory built into the electronic device. In other embodiments, the memory 11 may also be an external memory of the electronic device, that is, a memory externally connected to the electronic device.

[0175] In some embodiments, the memory 11 is used to store program codes and various data, and to achieve high-speed and automatic access to programs or data during the operation of the electronic device.

[0176] The memory 11 may include a random access memory and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0177] In one embodiment, the processor 12 may be a central processing unit (CPU), or 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, etc. A general-purpose processor may be a microprocessor or the processor may be any other conventional processor, etc. In one embodiment, the processor 12 may also be one or more of a CPU and a MCU (Microcontroller Unit). In this embodiment, the processor 12 includes the CPU 121 and the MCU 122. The CPU 121 is used to control and process all application programs of the electronic device 100. The CPU 121 includes at least one MCU 122. The MCU 122 is used to process the management program of the battery 13 to implement management functions, such as realizing the battery capacity estimation of the battery 13.

[0178] If the program code and various data in the memory 11 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, such as implementing the steps in the battery capacity estimation method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the MCU, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), etc. In one embodiment, the memory 11 is one or more of ROM, RAM (Random Access Memory, Random Access Device), and FLASH (Flash Memory). In this embodiment, the memory 11 may include ROM111, RAM112 and FLASH113. The ROM 111 may be used to store the algorithm program of the battery capacity estimation method. The ROM 111 may also store permanent data. The permanent data may include important information such as identity and password. The algorithm program and permanent data share the storage space in the ROM 111. The space in the ROM 111 is stored in a partition management manner. The RAM 112 may be used to store the functions and variables of the battery capacity estimation method. The FLASH 113 may be used to store the algorithm program, functions and variables of the battery capacity estimation method.

[0179] It is understandable that the module division described above is a logical function division, and there may be other division methods in actual implementation. In addition, the functional modules in each embodiment of the present application may be integrated in the same processing unit, or each module may exist physically separately, or two or more modules may be integrated in the same unit. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0180] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the spirit or basic features of the present application. Therefore, no matter from which point of view, the above embodiments of the present application should be regarded as exemplary and non-restrictive, and the scope of the present application is defined by the attached claims rather than the above description, and therefore it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present application.

Claims

1. A battery capacity estimation method, characterized in that: The method comprises: Obtain the charging current, charging voltage, charging time, and charging voltage range of the battery to be estimated; Determine a charging current interval and a charging time interval according to the charging current, the charging voltage and the charging time, wherein the charging current interval and the charging time interval correspond to the charging voltage interval; Determine the total training characteristics under different cycle numbers according to the charging current, charging voltage and charging time of the standard battery under different cycle numbers and the charging voltage range, wherein the total training characteristics include the total training electrochemical characteristics and the total training time domain characteristics; Determine a first total feature at different cycle numbers whose linear correlation between the total training feature at different cycle numbers and the actual capacity of the standard battery at different cycle numbers is greater than a preset value; Determining training features when training a model according to the first total features at different numbers of cycles, wherein the training features include training electrochemical features and training time domain features; Determine the actual characteristics of the battery to be estimated according to the charging voltage interval, the charging current interval, the charging time interval, and the training characteristics, wherein the actual characteristics include actual electrochemical characteristics and actual time domain characteristics; The actual capacity of the battery to be estimated is estimated according to the model and the actual characteristics.

2. The battery capacity estimation method according to claim 1, wherein: The model is a Gaussian process regression model.

3. The battery capacity estimation method according to claim 1, wherein: The difference between the maximum charging voltage and the minimum charging voltage in the charging voltage interval is less than or equal to 0.1 volt.

4. The battery capacity estimation method according to claim 1, wherein: The determining the actual characteristics of the battery to be estimated according to the charging voltage interval, the charging current interval, the charging time interval, and the training characteristics comprises: Determining the charging capacity of the battery to be estimated according to the charging current in the charging current interval and the charging time in the charging time interval; Determining a capacity increment value according to the charging voltage in the charging voltage interval and the charged capacity; The actual characteristics of the battery to be estimated are determined according to the charging current in the charging current interval, the charging voltage in the charging voltage interval, the charging time in the charging time interval, the capacity increment value, and the training characteristics.

5. The battery capacity estimation method according to claim 1, wherein: The total training electrochemical characteristics include activation energy increment, throughput and charging time, and the total training time domain characteristics include peak value, maximum value, mean value, variance, root mean square value, peak factor, kurtosis, kurtosis factor, waveform factor, pulse factor, margin factor, minimum value and skewness factor.

6. The battery capacity estimation method according to claim 1, characterized in that: Before determining the training feature according to the first total feature at different numbers of cycles, the method further includes: The total training features at different numbers of cycles are transformed into second total features at different numbers of cycles by using a principal component analysis method; Determining the training feature according to the first total feature at different numbers of cycles includes: Determine that the same feature in the first total feature at different cycle numbers and the second total feature at different cycle numbers is a third total feature at different cycle numbers; The training feature used when determining the training model is the third total feature.

7. The battery capacity estimation method according to claim 1, characterized in that: The method further comprises: Obtaining initial parameters generated when training the model using charging current, charging voltage, and charging time at different cycle times of a standard battery; Acquire a first feature when the initial parameter is generated, the first feature comprising a first electrochemical feature and a first time domain feature; Replacing the least relevant feature of the first feature with a second feature determined by the charging current, charging voltage and charging time at different cycle numbers in actual use of the battery to be estimated, wherein the second feature includes a second electrochemical feature and a second time domain feature; The trained model is obtained by updating the initial parameters through the updated first feature.

8. The battery capacity estimation method according to claim 7, It is characterized by: The first feature of obtaining and generating the initial parameters includes: Acquire a first feature at a first number of cycles when generating the initial parameters; The second feature determined by using the charging current, charging voltage and charging time at different cycle numbers in actual use of the estimated battery to replace the least relevant feature of the first feature includes: Determine a second characteristic at a second number of cycles according to the first characteristic, the charging current, the charging voltage, and the charging time at different cycle numbers in actual use of the battery to be estimated; Determine the smallest correlation at the second number of cycles among the correlations of the second feature at the second number of cycles and the first feature at the first number of cycles; The feature corresponding to the correlation at the second number of cycles in the first feature at the first number of cycles is replaced by the second feature at the second number of cycles.

9. An electronic device, characterized in that: The electronic device comprises: Battery; Processor; and A memory, wherein a plurality of program modules are stored in the memory, and the plurality of program modules are loaded by the processor and execute the battery capacity estimation method as described in any one of claims 1 to 8.

10. The electronic device as claimed in claim 9, wherein: The battery is a polymer lithium ion battery; The processor is one or more of a CPU and an MCU; The memory is one or more of ROM, RAM, and FLASH.

11. A storage medium having at least one computer instruction stored thereon, characterized in that: The instructions are loaded by the processor to execute the battery capacity estimation method as described in any one of claims 1-8.

Images