阅读说明：本技术 估计装置、估计方法以及计算机程序 (Estimation device, estimation method, and computer program ) 是由 鹈久森南 于 2020-05-21 设计创作，主要内容包括：本发明提供一种估计装置、估计方法以及计算机程序。估计装置(101)具备：获取部(201),获取蓄电元件的SOC的时间序列数据；分解部(202),将所述时间序列数据中的所述SOC的变动的波形分解成多个频率区域的波形分量；和估计部(204),基于各波形分量的强度和劣化系数来估计所述蓄电元件的劣化。(The invention provides an estimation device, an estimation method and a computer program. An estimation device (101) is provided with: an acquisition unit (201) that acquires time-series data of the SOC of the power storage element; a decomposition unit (202) which decomposes the waveform of the variation in SOC in the time-series data into waveform components in a plurality of frequency regions; and an estimation unit (204) that estimates the deterioration of the power storage element on the basis of the intensity of each waveform component and the deterioration coefficient.)

1. An estimation device is characterized by comprising:

an acquisition unit that acquires time-series data of the SOC of the power storage element;

a decomposition unit that decomposes a waveform of the variation in SOC in the time-series data into frequency components; and

an estimation unit that estimates deterioration of the power storage element based on the frequency component.

2. The estimation device according to claim 1,

the acquisition unit acquires time-series data of SOC for a predetermined period,

the decomposition unit processes a function of a temporal change in the SOC variation over the predetermined period.

3. The estimation device according to claim 1 or 2,

the decomposition unit decomposes a waveform of the variation in SOC in the time-series data into waveform components of a plurality of frequency regions,

the estimation portion estimates the degradation based on the intensity of each waveform component and a degradation coefficient.

4. The estimation device according to claim 3,

the disclosed device is provided with: a determination section that determines a degradation coefficient based on the waveform component,

the estimation portion estimates the degradation based on the degradation coefficient determined by the determination portion.

5. The estimation device according to claim 4,

the determination section determines the degradation coefficient based on a relationship of an intensity, a frequency, and a degradation coefficient of the waveform component.

6. The estimation device according to claim 4 or 5,

the determination section determines the degradation coefficient by weighting a waveform component having a larger intensity than a waveform component having a smaller intensity.

7. The estimation device according to claim 4,

the determination portion inputs the waveform component acquired by the decomposition portion to a learning model that outputs a degradation coefficient when the waveform component is input to determine the degradation coefficient.

8. An estimation method, comprising:

acquiring time-series data of the SOC of the electric storage element;

decomposing a waveform of the variation of the SOC in the time-series data into frequency components; and

estimating deterioration of the electrical storage element based on the frequency component.

9. A computer program for causing a computer to execute:

acquiring time-series data of the SOC of the electric storage element;

decomposing a waveform of the variation of the SOC in the time-series data into frequency components; and

estimating deterioration of the electrical storage element based on the frequency component.

Technical Field

The present invention relates to an estimation device, an estimation method, and a computer program for estimating deterioration of a power storage element.

Background

An electric storage device capable of storing electric energy and supplying energy as a power source when necessary is being used. The power storage element is applied to portable devices, power supply devices, transportation equipment including automobiles and railways, industrial equipment including aviation, space, and construction, and the like. It is important to always grasp the storage capacity of the storage element so that the required amount of energy stored can be used when necessary. It is known that an electric storage element is mainly chemically deteriorated depending on time and frequency of use. Therefore, the energy that can be used decreases with time and frequency of use. In order to utilize the required amount of energy when necessary, it is important to grasp the state of degradation of the power storage element. Heretofore, techniques for estimating the deterioration of the power storage element have been developed.

There is a demand for a technique that can estimate the SOH (State Of Health) well even if the SOC (State Of Charge) Of the power storage element varies variously.

The present inventors have found that, as shown in patent document 1, when the fluctuation range of the SOC around a given SOC is large, the degradation value is large, and even if the fluctuation range of the SOC is the same, the degradation value varies depending on the central SOC. An estimation device that estimates degradation of the power storage element based on the magnitude of the variation in SOC and the SOC of the center has been developed. In the estimation device of patent document 1, the degradation value is estimated to increase according to the magnitude of the variation in SOC.

Prior art documents

Patent document

Patent document 1: japanese patent No. 6428957

Disclosure of Invention

Problems to be solved by the invention

In the conventional method, the accuracy of estimating the deterioration of the power storage element may be insufficient.

Even if the fluctuation range of the SOC is small, the SOC may deteriorate when charged at a high rate. It is required that deterioration can be estimated well even in a complicated SOC fluctuation pattern or a fluctuation pattern with a small SOC fluctuation width.

An object of the present invention is to provide an estimation device, an estimation method, and a computer program that can estimate the deterioration of a power storage element with high accuracy.

Means for solving the problems

An estimation device according to an aspect of the present invention includes: an acquisition unit that acquires time-series data of the SOC of the power storage element; a decomposition unit that decomposes a waveform of the variation in SOC in the time-series data into frequency components; and an estimation unit that estimates deterioration of the power storage element based on the frequency component.

An estimation method according to an aspect of the present invention acquires time-series data of an SOC of an electric storage element, decomposes a waveform of a fluctuation of the SOC in the time-series data into frequency components, and estimates degradation of the electric storage element based on the frequency components.

A computer program according to an embodiment of the present invention causes a computer to execute: time-series data of an SOC of an electric storage element is acquired, a waveform of variation of the SOC in the time-series data is decomposed into frequency components, and deterioration of the electric storage element is estimated based on the frequency components.

Effects of the invention

In the present invention, the deterioration of the power storage element can be estimated with high accuracy.

Drawings

Fig. 1 is a diagram showing an example of a change in the deterioration value of the power storage element due to energization with respect to the fluctuation range of the SOC.

Fig. 2 is a diagram showing an example of a change in the deterioration value of the power storage element due to energization with respect to the center SOC.

Fig. 3 is a graph showing the relationship between time and SOC when the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, and Δ 30%.

Fig. 4A is a graph showing waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 3.

Fig. 4B is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 3.

Fig. 4C is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 3.

Fig. 4D is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 3.

Fig. 4E is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 3.

Fig. 5 is a graph showing a relationship between time and SOC when the SOC is varied by Δ 1.5% and Δ 20%.

Fig. 6 is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 5.

Fig. 7 is a diagram showing the structure of the monitoring device.

Fig. 8 is a diagram showing the structure of the estimation device.

Fig. 9 is a flowchart showing the procedure of a process in which the estimation device performs calculation of a degradation value based on time-series data.

Fig. 10 is a graph showing the relationship between time and SOC when the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, and Δ 30%.

Fig. 11 is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 10.

Fig. 12 is a graph showing the relationship between the number of days and cycle deterioration when the SOC was varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, Δ 30%, Δ 100%.

Fig. 13 is a graph showing the relationship of the amplitude spectrum, the frequency, and the degradation coefficient.

Fig. 14 is a graph showing the relationship between the amplitude spectrum, the frequency, and the degradation coefficient when interpolation is performed by interpolation calculation based on k in each graph of fig. 12.

Fig. 15 is a graph showing a relationship between the estimated value and the measured value.

Fig. 16 is a graph showing time-series data of the SOC.

Fig. 17 is a block diagram showing an example of the configuration of the estimation device according to embodiment 2.

Fig. 18 is an explanatory diagram showing an example of the record layout of the training data DB.

Fig. 19 is a diagram showing the structure of the degradation coefficient output model.

Fig. 20 is a flowchart showing an example of a processing procedure of the deterioration coefficient output model generation processing performed by the control unit.

Fig. 21 is a flowchart of an operation procedure when the estimation device determines to estimate the deterioration of the power storage element.

Detailed Description

(outline of embodiment)

An estimation device according to an embodiment includes: an acquisition unit that acquires time-series data of the SOC of the power storage element; a decomposition unit that decomposes a waveform of the variation in SOC in the time-series data into frequency components; and an estimation unit that estimates deterioration of the power storage element based on the frequency component.

According to the above configuration, since the waveform of the fluctuation of the SOC is decomposed into frequency components, it is possible to detect a waveform having a large fluctuation but a long period (fluctuation time) and a waveform having a small fluctuation but a short period. A waveform having a large variation but a long period becomes a waveform component (frequency component) having a large intensity (spectrum intensity, amplitude spectrum in the case of fourier transform) and a low frequency. A waveform with small fluctuation but short period becomes a waveform component with small intensity and high frequency. If the cycle is very long even if the fluctuation is large, the deterioration is small. If the cycle is very short even if the fluctuation is small, the deterioration is large. When the electric storage element is a lithium ion secondary battery, lithium deposition or the like may occur in the negative electrode when rapid charging is performed at a high frequency on the high SOC side, resulting in a large deterioration. The waveform of the variation in SOC differs depending on the characteristics of the power storage element and the usage pattern of the user, but according to the above configuration, it is possible to detect an arbitrary waveform and to estimate the deterioration of the power storage element satisfactorily.

In the above-described estimation device, the acquisition unit may acquire time-series data of SOC for a predetermined period, and the decomposition unit may process a function of a time change of SOC fluctuation for the predetermined period.

According to the above configuration, the temporal change in SOC variation over a predetermined period is acquired, and the temporal change in SOC variation is analyzed. The degradation estimation is not performed from time to time based on the data of SOC fluctuation, but performed in batch processing (batch processing) by acquiring the temporal change of SOC fluctuation over a predetermined period. In a time change of a certain time width of the SOC fluctuation range, it is possible to analyze what kind of SOC change pushes degradation, and to perform degradation estimation satisfactorily.

In the above-described estimation device, the decomposition unit may decompose a waveform of the variation in the SOC in the time-series data into waveform components of a plurality of frequency regions, and the estimation unit may estimate the degradation based on an intensity of each waveform component and a degradation coefficient.

According to the above configuration, the degradation is estimated based on the magnitude of the variation in SOC and the degradation coefficient. Deterioration can be estimated well by taking into account the fact that the amount of change in SOH increases in accordance with the fluctuation range of SOC, and also taking into account a waveform having a very small fluctuation range based on an actual complex fluctuation pattern.

The above estimation device may further include: a determination section that determines a degradation coefficient based on the waveform component, the estimation section estimating the degradation based on the degradation coefficient determined by the determination section.

According to the above configuration, the degradation coefficient used for the calculation of the degradation is determined based on the waveform component in accordance with the SOC fluctuation width and the SOC center. The deterioration coefficient is determined from the distribution of the waveform component in the frequency region, the intensity and frequency of the waveform component of the high peak, and the like, and the deterioration can be estimated well and easily.

In the above-described estimation device, the determination unit may determine the degradation coefficient based on a relationship between the intensity and frequency of the waveform component and the degradation coefficient.

According to the above configuration, the relationship between the intensity and frequency of the waveform component and the degradation coefficient is obtained in advance, and the degradation coefficient is determined based on the relationship, whereby the degradation can be estimated favorably and easily.

In the above-described estimation device, the determination unit may determine the degradation coefficient by weighting a waveform component having a larger intensity than a waveform component having a smaller intensity.

Since weighting is performed based on the distribution of the waveform component in the frequency region, the intensity and frequency of the waveform component of the high peak, and the like, and the degradation coefficient is determined based on the amplitude and frequency of the waveform component obtained by the weighting, it is possible to estimate the degradation favorably.

In the above-described estimation device, the determination unit may input the waveform component acquired by the decomposition unit to a learning model that outputs the degradation coefficient when the waveform component is input, and determine the degradation coefficient.

According to the above configuration, the deterioration coefficient is determined by taking into account the waveform component having a small spectral intensity and a large frequency, and the deterioration can be estimated favorably.

An estimation method according to an embodiment acquires time-series data of an SOC of an electric storage element, decomposes a waveform of a variation of the SOC in the time-series data into frequency components, and estimates degradation of the electric storage element based on the frequency components. The waveform of the variation in SOC in the time-series data may be decomposed into waveform components in a plurality of frequency regions, and the deterioration of the power storage element may be estimated based on the intensity and the deterioration coefficient of each waveform component.

According to the above configuration, since the waveform of the variation in SOC is decomposed, it is possible to detect a waveform having a long period and a waveform having a short period and a large variation. Any waveform can be detected according to the characteristics of the power storage element and the usage pattern of the user, and the deterioration of the power storage element can be estimated favorably.

The computer program according to the embodiment causes a computer to execute: time-series data of an SOC of an electric storage element is acquired, a waveform of variation of the SOC in the time-series data is decomposed into frequency components, and deterioration of the electric storage element is estimated based on the frequency components. The computer may also be caused to execute: the waveform of the variation in the SOC in the time-series data is decomposed into waveform components of a plurality of frequency regions, and the deterioration of the power storage element is estimated based on the intensity and the deterioration coefficient of each waveform component.

Fig. 1 is a diagram illustrating an example of a change in the amount of deterioration of the power storage element due to energization with respect to the fluctuation range of the SOC. In fig. 1, the ordinate represents the difference between the degradation amount when a predetermined electric energy is applied and the degradation amount at the SOC variation range of 3%, and the abscissa represents the SOC variation range.

In fig. 1, the deterioration amount due to energization after repeating charge and discharge a given number of times so that the center SOC becomes 60% is plotted against the fluctuation width of the SOC.

As shown in fig. 1, the present inventors have found that the deterioration amount due to energization varies when the SOC variation width varies even though the center SOC is the same. It was found that the deterioration due to the energization increased in accordance with the magnitude of the fluctuation in SOC.

The mechanism of this phenomenon has not yet been fully elucidated. The inventors have considered that as the magnitude of the variation in SOC increases, the expansion (during charging) and contraction (during discharging) of the negative electrode become more significant, and the SEI film formed on the surface of the negative electrode is partially broken, and as a result, the amount of deterioration of the power storage element due to energization increases.

Fig. 2 is a diagram illustrating an example of a change in the deterioration amount of the power storage element due to energization with respect to the center SOC. In fig. 2, the vertical axis represents the difference between the degradation amount in the case where a given amount of electricity is applied and the degradation amount at the center SOC of 10%, and the horizontal axis represents the center SOC which is the center of the variation in SOC. Here, the center SOC means the center of variation of the SOC in the time-series data of the SOC.

In fig. 2, the deterioration amount due to energization after the SOC fluctuation range becomes 10% after repeating charge and discharge a given number of times is plotted against the center SOC.

Here, a description will be given of an example of the charge/discharge operation. Repeating charge and discharge so that the central SOC becomes 10% and the fluctuation range of the SOC becomes 10% means repeating charge and discharge so that the SOC reciprocates between 0% and 20%. It is found that the deterioration amount greatly differs depending on the center SOC even if the variation amount is the same.

As can be seen from the above, it is necessary to estimate the SOH in consideration of not only the total variation amount of the SOC but also the variation width and the center SOC of the SOC.

As shown in patent document 1, the present inventors have focused on the SOC fluctuation range and developed an estimation device that estimates the deterioration of the power storage element based on the magnitude of the SOC fluctuation and the center SOC. As described above, in patent document 1, the degradation value is increased according to the magnitude of the SOC fluctuation, and the estimation accuracy is improved. It is desirable to detect a waveform having a very small SOC fluctuation width and to estimate a degradation value more favorably with respect to a complicated SOC fluctuation pattern.

A waveform of variation in SOC for a given acquisition period is decomposed into waveform components of a plurality of frequency regions, and deterioration of the power storage element is estimated based on the intensity of the waveform components and a deterioration coefficient. The present inventors have found that degradation can be estimated more favorably by detecting a waveform having a plurality of fluctuation ranges including a waveform having a very small fluctuation range, simulating an actual complex fluctuation pattern, while taking into consideration the fact that the amount of change in SOH increases in accordance with the fluctuation range of SOC by decomposing the waveform into waveform components, and have completed the present invention.

The present invention can be expressed as follows. The inventors have found that accuracy of degradation estimation may be improved by acquiring a temporal change in SOC variation over a certain time width (a certain period) and analyzing the temporal change in SOC variation. That is, the deterioration estimation is not performed from time to time based on the SOC variation data, but performed in a batch process (batch process) by acquiring a time variation of SOC variation of a certain time width (a certain period). The key point of the present invention is to analyze what kind of SOC change pushes degradation in a time change of a SOC fluctuation range of a certain time width, and to estimate degradation. Therefore, in the temporal change of the SOC fluctuation range within a certain period, the temporal change of the SOC fluctuation is expressed by another variable, and the temporal change of the original SOC fluctuation range is expressed by another function. As this method, a frequency conversion method is generally adopted, and as a widely known method, fourier transform is exemplified. A Fast Fourier Transform (FFT) is typically performed. In addition, when the temporal change in SOC variation of a certain time width (certain period) is expressed by another function, it is optimal to express the temporal change in SOC variation by a plurality of (many) functions, but since it takes time for analysis (control), the temporal change in SOC variation may be expressed by only a function that contributes most to the temporal change in SOC variation.

Alternatively, the averaging process may be performed for a temporal change in SOC variation over a certain time width (a certain period).

Fig. 3 is a graph showing the relationship between time and SOC when the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, and Δ 30%. The horizontal axis represents time [ sec ], and the vertical axis represents SOC [% ]. Graphs of Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, Δ 30% are represented by a, b, c, d, e.

Fig. 4A to 4E are graphs of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 3. The horizontal axis is frequency [ Hz ], and the vertical axis is amplitude (amplitude spectrum) [% ]. The complex waveform can be decomposed into a plurality of sine (sin) waveforms by fourier transform, and the obtained f (ω) can be represented by a graph of the magnitude (amplitude) of the sine (sin) wave having an angular frequency on the horizontal axis and angular frequencies ω on the vertical axis. In fig. 4A to 4E, angular frequencies are replaced with frequencies. The fluctuation range of the waveform of fig. 3 relates to the amplitude of the waveform components of fig. 4A to 4E, and the period of the waveform of fig. 3 corresponds to the center of gravity of the frequency of some of the waveform components of fig. 4A to 4E. The complex waveform of fig. 3 is frequency transformed into f (ω) by fourier transform, and the contribution of f (ω) is represented as spectral intensity (peak height or peak area). Time-series data over a long period can be estimated by acquiring time-series data of an SOC over a short acquisition period. The usage pattern of the power storage element by the user is also almost fixed in many cases.

The graphs of fig. 4A to 4E correspond to the waveforms of a, b, c, d, and E of fig. 3, respectively.

As shown in fig. 4A to 4E, the waveforms of a, b, c, d, and E of fig. 3 are decomposed into waveform components of a plurality of frequency regions, respectively. The frequency of the peak component among the waveform components corresponding to the waveform of a is approximately 0.013, and the amplitude is approximately 0.133, which is shown in fig. 4A. The peak component of the waveform component of fig. 4E corresponding to the waveform E of fig. 3 and Δ 30% has a large amplitude due to a large variation in the waveform E of fig. 3, and a small frequency due to a long period. The peak component of the waveform component in fig. 4B corresponding to the waveform B of Δ 1.5% has a small amplitude because of a small fluctuation of the waveform B in fig. 3, and has a large frequency because of a short cycle.

Fig. 5 is a graph showing a relationship between time and SOC when the SOC is varied by Δ 1.5% and Δ 20%. The horizontal axis represents time [ sec ], and the vertical axis represents SOC [% ]. The graphs of Δ 1.5%, Δ 20%, and the sum of the graphs of Δ 1.5% and Δ 20% are represented by a, b, and c.

Fig. 6 is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 5. The horizontal axis is frequency (Hz) and the vertical axis is amplitude. As is clear from fig. 6, most of the waveform components correspond to the waveform b of Δ 20%, and the waveform components include the waveform component corresponding to the waveform a of Δ 1.5%. The degradation can be estimated based on the waveform component having a large amplitude. That is, it is possible to estimate the degradation by weighting based on a waveform component having a large amplitude.

The estimation method according to the present embodiment acquires time-series data of the SOC of the power storage element, and decomposes a waveform of variation in the SOC in the time-series data into waveform components in a plurality of frequency regions. The waveform component is weighted for a waveform component having a larger amplitude than a waveform component having a smaller intensity (amplitude). The relationship between the amplitude, frequency and degradation coefficient is obtained in advance. The degradation coefficient k is determined based on the amplitude and frequency of the emphasized waveform component obtained by weighting. Based on the determined degradation coefficient k, a degradation value (cyclic degradation) is calculated by the following equation (1).

Here, t is an elapsed time. The degradation value increases, for example, according to root law. Increasing according to the root law means that the increment per unit time of the degradation value gradually decreases with the passage of time (see fig. 15).

(embodiment mode 1)

Hereinafter, the description will be specifically made based on the drawings showing the embodiments.

Fig. 7 is a diagram showing the structure of the monitoring device 151. The monitoring device 151 includes an a/D conversion unit 53, a history creation unit 54, a counter 55, a storage unit 56, a communication unit 57, and an estimation device 101. The monitoring device 151 is connected to a current sensor 41, a voltage sensor 42, and a temperature sensor 43.

Some of the components included in the monitoring device 151 may be disposed separately from the other components. For example, the estimation device 101 may be remotely disposed and communicate with the communication unit 57. A server that is remotely located and connected to a network may also function as the estimation device 101.

The monitoring device 151 monitors deterioration of an electric storage element (lithium ion 2-time battery in the present embodiment) to be monitored. The monitoring device 151 may be configured to monitor 1 battery cell, or may be configured to monitor a plurality of battery cells (battery packs) connected in series or in parallel. The monitoring device 151 may constitute an electric storage device (battery pack) together with the battery pack.

The monitoring target storage element is not limited to a nonaqueous electrolyte 2-time cell such as a lithium ion 2-time cell, and may be another electrochemical cell suitable for assumptions, algorithms, and mathematical models described later. Hereinafter, the power storage element to be monitored is simply referred to as a battery.

The counter 55 in the monitoring device 151 counts clock pulses generated by an oscillation circuit using a crystal oscillator or the like, and holds the counted value. The count value may also indicate the current time.

The current sensor 41 measures a current charged into and discharged from the electric storage element, and outputs an analog signal Ai representing the measurement result to the a/D conversion unit 53.

The voltage sensor 42 measures the voltage between the positive electrode and the negative electrode of the power storage element, and outputs an analog signal Av indicating the measurement result to the a/D converter 53.

The temperature sensor 43 measures the temperature T of a predetermined portion of the storage element, and outputs an analog signal At indicating the measurement result to the a/D conversion unit 53.

The a/D conversion unit 53 converts the analog signals Ai, Av, and At received from the current sensor 41, the voltage sensor 42, and the temperature sensor 43, respectively, into digital signals Di, Dv, and Dt, for example, At each predetermined sampling time.

The history creation unit 54 stores the count value of the counter 55 at the sampling time and the digital signals Di, Dv, and Dt in the storage unit 56. The storage unit 56 stores the sampling time, the current value, the voltage value, and the temperature T for each sampling time. The number of charge/discharge cycles is stored in the storage unit 56, and is updated every time charge/discharge is repeated.

The communication Unit 57 may communicate with other devices such as a main Control device (Electronic Control Unit) in the vehicle, a personal computer, a server, a smartphone, and a terminal for maintaining the power storage element.

The communication section 57, if receiving an estimation command of, for example, the deterioration state of the power storage element from another device, outputs the received estimation command to the estimation device 101.

Fig. 8 is a diagram showing the configuration of the estimation device 101.

Referring to fig. 8, the estimation device 101 includes a control unit 20, a storage unit 23, and an interface unit 24. The interface unit 24 is configured by, for example, a LAN interface, a USB interface, or the like, and performs communication with another device such as the monitoring device 151 by wire or wireless.

The signal line or terminal from the estimation device 101 to the communication unit 57 may also function as an output unit that outputs the estimation result or the like. The communication unit 57 may function as an output unit.

If different input data is input to the estimation device 101, different outputs are obtained from the output section. When different SOC fluctuation ranges (and/or center SOCs) are input to the estimation device 101, the output unit may output different outputs (for example, voltage values and duty ratios).

A display unit (or a notification unit) for displaying the output result may be connected to the output unit. The display unit (or the notification unit) may be caused to display and output from the output unit via the communication unit 57.

A degradation estimation program 231 for executing a degradation estimation process described later is stored in the storage unit 23. The degradation estimation program 231 is provided in a state of being stored in the computer-readable recording medium 60 such as a CD-ROM, a DVD-ROM, or a USB memory, and is stored in the storage unit 23 by being installed in the estimation device 101. The degradation estimation program 231 may be acquired from an external computer, not shown, connected to the communication network and stored in the storage unit 23.

The storage unit 23 stores a degradation coefficient DB 232 required for the degradation estimation process. The degradation coefficient DB 232 stores the relationship between the amplitude and frequency of the waveform component and the degradation coefficient for each of the center SOC, the temperature, and the acquisition period of the time-series data of the acquisition SOC, for example. In addition, when any one of the center SOC, the temperature, and the acquisition period is fixed, for example, when the acquisition period is fixed, it is not necessary to store the relationship for each acquisition period. For example, the temporal change of the degradation value when the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, Δ 30% for each temperature T and for each SOC center is measured by a previous experiment. The deterioration coefficient k is calculated based on the measurement result of the test.

The control unit 20 is configured by, for example, a CPU, a ROM, a RAM, or the like, and controls the operation of the estimation device 101 by executing a computer program such as the degradation estimation program 231 read from the storage unit 23. The control unit 20 reads and executes the degradation estimation program 231, thereby functioning as a processing unit that executes degradation estimation processing.

The control section 20 includes an acquisition section 201, a decomposition section 202, a determination section 203, and an estimation section 204.

The acquisition unit 201 acquires time-series data of the SOC of the power storage element. More specifically, if an estimation command is received from the communication unit 57, the acquisition unit 21 acquires each sampling time and the current value, the voltage value, and the temperature T at each sampling time from the storage unit 56 in the monitoring device 151 via the interface unit 24 in accordance with the received estimation command. The sampling interval and the acquisition period for sampling are determined according to the characteristics of the electric storage element, the usage method, and the like. If the sampling interval is short, the waveform is sometimes interrupted. When the fluctuation width is large and the period is long, the acquisition period needs to be extended.

In this manner, the acquisition unit 21 acquires data measured after the use of the power storage element is started from the storage unit 56.

Alternatively, the acquisition section 21 may acquire data from a data file.

The acquisition unit 21 secures a storage area for storing data for the sampling time, SOC, and temperature.

The acquisition unit 21 calculates the amount of electric power to be supplied to the power storage element by, for example, summing up the current values at each sampling time, and converts the calculated amount of electric power into the amount of change in SOC. Then, the acquisition unit 21 calculates the SOC at each sampling time based on the conversion result. The acquisition unit 21 may correct the SOC using a measured value of the open circuit voltage, for example.

The decomposition unit 202 performs fourier transform on the waveform of the fluctuation of the SOC in the time-series data of the SOC acquired by the acquisition unit 21, and decomposes the waveform into waveform components in a plurality of frequency regions.

The determination unit 203 weights a waveform component having a larger amplitude than a waveform component having a smaller amplitude. The weighting is performed based on the peak height of each waveform component, the area of each waveform component, or the like. The determination section 203 determines the degradation coefficient k based on the amplitude and frequency of the emphasized waveform component obtained by weighting.

The estimation unit 204 calculates a degradation value, for example, a cycle degradation (a decrease amount of capacity) by the above equation (1) based on the degradation coefficient k determined by the determination unit 203.

The estimation unit 204 may transmit estimation result information indicating the calculated degradation value to another device via the communication unit 57 as a response to the estimation command.

The estimation unit 204 can estimate the deterioration of the power storage element due to the energization from the change in the state of the coating film in the electrode based on the magnitude of the SOC variation, for example. The growth rate of the SEI film decreases as the SEI film is formed, but when the variation in SOC is large and the SEI film is partially broken down, the growth rate of the SEI film increases and the degradation value increases. In the present embodiment, a fluctuation waveform having a large fluctuation width is taken into account by weighting, and a fluctuation waveform having a small fluctuation width is also detected by fourier transform, so that the accuracy of the estimation of the deterioration is high.

The monitoring device 151 or the estimation device 101 of the monitoring device 151 includes the control unit 20, and the control unit 20 reads from the storage unit 23 and executes the degradation estimation program 231 including a part or all of the steps of the flowcharts shown below.

Fig. 9 is a flowchart of the operation steps when the estimation device determines to estimate the deterioration of the power storage element.

Referring to fig. 9, a situation is assumed in which the control unit 20 of the estimation device 101 receives an estimation command from another device. Hereinafter, for example, a case where charging and discharging are repeated so that the center SOC is 60% and the SOC is in a range of 30% to 90% will be described.

First, the control unit 20 acquires SOC variation data for a predetermined period (S1).

The control unit 20 performs fourier transform on the waveform of the fluctuation in SOC to obtain f (ω), and converts the waveform into a function f (f) of frequency (S2).

The control unit 20 performs weighting (S3). Weighting is performed based on the peak height of each waveform component, the area (integrated value) of each waveform component, or the like.

The control portion 20 determines the degradation coefficient k (S4). The control unit 20 reads the degradation coefficient DB 232 and acquires the relationship between the amplitude, frequency, and degradation coefficient according to the center SOC, temperature, and acquisition period. The control unit 20 determines the degradation coefficient kr from the relationship among the amplitude, the frequency, and the degradation coefficient based on the amplitude and the frequency of the weighted waveform component.

The control unit 20 calculates Δ SOH by the above equation (1) based on the determined degradation coefficient k (S5), and ends the process.

The following description will be specifically made.

Fig. 10 is a graph showing the relationship between time and SOC when the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, and Δ 30%. The horizontal axis represents time [ sec ], and the vertical axis represents SOC [% ]. In fig. 10, a graph in which the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, and Δ 30% is shown by a, b, c, d, and e. Fig. 10 also shows the variation in SOC when the power storage element is actually used (example).

Fig. 11 is a graph of waveform components decomposed into a plurality of frequency regions by fourier transform of the waveform of fig. 10. The horizontal axis represents frequency [ Hz ], and the vertical axis represents amplitude spectrum [% ].

Similarly to fig. 6, waveform components corresponding to the waveforms b, c, d, and e in fig. 10 are shown, and waveform components obtained by decomposing the waveform of the embodiment are also shown. The peak component among the waveform components corresponding to the waveform of a in fig. 10 has a frequency of substantially 0.013 and an amplitude of substantially 0.133, which is not shown in fig. 10.

The degradation coefficient k of each Δ SOC is obtained experimentally as described above.

Fig. 12 is a graph showing the relationship between the number of days and cycle deterioration when the SOC was varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, Δ 30%, Δ 100%. In FIG. 12, the horizontal axis represents day [ 2 ]]The vertical axis is the cyclic deterioration [% ]]. The cycle deterioration is represented by a standard value when the maximum deterioration amount in fig. 12 is 100%. Hereinafter, fig. 13 to 15 are also based on this normalization. Fig. 12 shows graphs in which the SOC is varied by Δ 0.5%, Δ 1.5%, Δ 5%, Δ 20%, Δ 30%, and Δ 100% using a, b, c, d, e, and f.

According to the formula (1),k for each fluctuation range is obtained by optimization calculation or the like based on the curves a to f.

Fig. 13 is a graph showing the relationship of the amplitude spectrum, the frequency, and the degradation coefficient. In FIG. 13, the x-axis is the amplitude spectrum [% ]]The y-axis being the frequency [ Hz]And the z-axis is the deterioration coefficient [% ]]. The relationship is expressed by associating the frequency and intensity of the peak top of each waveform of a, b, c, d, and e in fig. 11 with k in each graph of a, b, c, d, and e obtained in fig. 12. In fig. 13, the amplitude spectrum is about 10.2% and the amplitude is about 2 × 10^-4In the case of (3), the degradation coefficient k is approximately 3.56.

Fig. 14 shows the relationship between the amplitude spectrum, the frequency, and the degradation coefficient when interpolation is performed by interpolation calculation based on k in each graph of fig. 12. In the case of fig. 14, k can be obtained by reading the value on the z-axis corresponding to the amplitude spectrum and frequency of the waveform other than the waveforms a, b, c, d, and e in fig. 11.

As described above, the control unit 20 weights the waveforms of the embodiment of fig. 11.

In fig. 14, the deterioration coefficient k is obtained by reading the z-coordinate of the read point o, which is a value on the z-axis corresponding to the amplitude spectrum and frequency of the emphasized waveform obtained by weighting. Here, k is 0.75.

Fig. 15 is a graph showing a relationship between the estimated value and the measured value. In fig. 15, the horizontal axis represents total days [ days ], and the vertical axis represents cycle deterioration [% ]. The estimated value represents the relationship between the total number of days and cycle deterioration when k is 0.1875 in expression (1).

The experimental values are plotted for the cycle deterioration at each time point with respect to the total number of days at the plurality of measurement time points. As can be seen from fig. 15, the estimation method according to the present embodiment can perform favorable estimation.

As shown in fig. 16, according to the estimation method of the embodiment, even when charge and discharge are repeated on the high SOC side and the amplitude of the fluctuation of SOC is biased to the positive (Plus) side, the deterioration value can be estimated favorably by decomposing into a plurality of waveform components by fourier transform and acquiring the deterioration coefficient k.

As described above, it is confirmed that the waveform of the fluctuation of the SOC is decomposed into waveform components of a plurality of frequency ranges, and the deterioration of the power storage element can be estimated with high accuracy based on each waveform component and the deterioration coefficient k.

By estimating the amount of decrease (the value of capacity degradation) in the amount of electricity reversibly taken out from the electricity storage element, the internal state of the electricity storage element can be grasped. Since the potential at which the SOC of the negative electrode is 100% is also known, the risk of deposition of metallic lithium in the negative electrode is also known when the storage element is a lithium ion secondary battery. The SOH of the power storage element including this risk can be monitored. It is possible to determine how to control the storage element.

(embodiment mode 2)

Fig. 17 is a block diagram showing an example of the configuration of the estimation device 101 according to embodiment 2. In the drawings, the same components as those in fig. 8 are denoted by the same reference numerals, and detailed description thereof is omitted.

The storage unit 23 of the estimation device 101 according to embodiment 2 has the same configuration as the estimation device 101 according to embodiment 1, except that the training data DB233 and the degradation coefficient output model 234 are stored.

Fig. 18 is an explanatory diagram showing an example of the record layout of the training data DB 233. The training data DB233 stores a plurality of pieces of training data for generating the degradation coefficient output model 234.

The training data DB233 stores a No. (number) column, a graph column after fourier transform, and a degradation coefficient k column. The No. column stores No. for identifying each training data. The graph sequence after the fourier transform stores graphs after the fourier transform of the time-series data of charge and discharge. The graph is a graph of f (f) obtained by converting ω of f (ω) into f (frequency (Hz)). The degradation coefficient k column stores a degradation coefficient k corresponding to the graph. For example, when the graph of fig. 11 obtained by fourier transform of the graph of the example of fig. 10 is used as training data, the degradation coefficient k obtained from the data of the experimental values of fig. 15 is stored in the degradation coefficient k column.

Fig. 19 is a diagram showing the structure of the degradation coefficient output model 234.

The degradation coefficient output model 234 is assumed to be a learning model used as a program module that is a part of artificial intelligence software and that can use a multilayer Neural Network (deep learning), for example, a Convolutional Neural Network (CNN), or a Recursive Neural Network (RNN). Other machine learning such as decision trees, random forests, support vector machines, etc. may also be used. The control unit 20 operates as follows: in accordance with the command from the degradation coefficient output model 234, a calculation is performed on the graph after fourier transform of the time-series data of charge and discharge of the input layer input to the degradation coefficient output model 234, and the probability value of the degradation coefficient k is output. The degradation coefficient output model 234 will be described in detail below.

The control unit 20 constructs a neural network that receives the acquired graph f (f) as an input and outputs information indicating the degradation coefficient k by learning the feature values as the degradation coefficient output model 234. The neural network is a CNN, and has an input layer that accepts input of the graph f (f), an output layer that outputs a probability value of the degradation coefficient k, and an intermediate layer that extracts feature quantities.

In fig. 19, the number of intermediate layers is set to 3, but the present invention is not limited thereto. For example, when the degradation coefficient output model 234 is CNN, the intermediate layer has a structure in which the convolutional layer and the pooling layer are alternately connected, and finally extracts the feature amount while compressing the information. In fig. 19, the convolutional layer and the pooling layer are not described. The output layer has a plurality of neurons that output the determination result of the degradation coefficient k, and outputs the plurality of degradation coefficients k and the probability values thereof based on the feature quantity output from the intermediate layer. Specifically, in the case of using a flexible maximum value transfer (Softmax) function, the deterioration coefficients k output from the neuron are a plurality of deterioration coefficients k and probability values thereof.

Further, although the degradation coefficient output model 234 is described as a model of CNN, RNN can be used as described above. In RNN, the intermediate layer at the previous time is used for learning together with the input layer at the next time.

The control unit 20 performs learning by using training data in which a graph of each No. of the training data DB233 after fourier transform and a degradation coefficient k in each graph are associated with each other.

The control unit 20 inputs the graph of f (f) as training data to the input layer, and acquires the probability value of the degradation coefficient from the output layer through calculation processing in the intermediate layer.

The control unit 20 compares the determination result output from the output layer with the correct value, which is information obtained by labeling the graph of f (f) in the training data, and optimizes the parameters used for the calculation processing in the intermediate layer so that the output value from the output layer approaches the correct value. The parameter is, for example, a weight (coupling coefficient) between neurons, a coefficient of an activation function used for each neuron, or the like. The method of optimizing the parameters is not particularly limited, and the control unit 20 optimizes various parameters by using an error back propagation method, for example.

The control unit 20 performs the above-described processing on the graph f (f) of each piece of training data included in the training data DB233, and generates the degradation coefficient output model 234. When acquiring the graph of f (f) obtained by fourier transforming the time-series data of charge and discharge, the control unit 20 acquires the degradation coefficient k showing a high probability value as the determined degradation coefficient k based on the probability value of the degradation coefficient k output from the degradation coefficient output model 234 and the threshold value using the degradation coefficient output model 234.

The control unit 20 can relearn the degradation coefficient output model 234 based on the degradation coefficient k determined by the degradation coefficient output model 234 and the actually measured degradation coefficient k so as to improve the reliability of the determination result.

Fig. 20 is a flowchart showing an example of processing procedures of the generation processing of the degradation coefficient output model 234 by the control unit 20.

The control unit 20 reads the training data DB233, and acquires training data in which the graph after fourier transform of each No. and the degradation coefficient are associated with each other (step S11).

The control unit 20 generates a degradation coefficient output model 234 that outputs the determined degradation coefficient k when the graph f (f) after fourier transform is input, using the training data (step S12). The control unit 20 stores the generated degradation coefficient output model 234 in the storage unit 23, and ends the series of processing.

Fig. 21 is a flowchart of the operation procedure when the estimation device 101 determines to estimate the deterioration of the power storage element.

First, the control unit 20 acquires SOC variation data for a predetermined period (S21).

The control unit 20 performs fourier transform on the waveform of the fluctuation in SOC to obtain f (ω), and converts the waveform into a function f (f) of frequency (S22).

The control unit 20 inputs the graph of f (f) to the degradation coefficient output model 234 (S23).

The control portion 20 determines the degradation coefficient k (S24). The control unit 20 sets the deterioration coefficient k having a probability value equal to or greater than a threshold value as the determined deterioration coefficient k based on the deterioration coefficient k and the probability value output from the deterioration coefficient output model 234.

The control unit 20 calculates Δ SOH by the above equation (1) based on the determined degradation coefficient k (S25), and ends the process.

According to the present embodiment, the deterioration coefficient is determined taking into account waveform components having small spectral intensities and large frequencies, and deterioration can be estimated favorably.

In addition, a plurality of degradation coefficient output models 234 may be generated according to the center SOC, the temperature, and the like.

The degradation coefficient output model may be configured to output the degradation coefficient k by inputting the time-series data of charge and discharge without inputting the graph of f (f).

In embodiments 1 and 2, the estimation device 101 uses time-series data of SOC, but the time-series data of SOC may be Δ SOC obtained by a current accumulation method or the like, or may be data obtained by adding/subtracting Δ SOC to/from an initial value of SOC.

In the estimation device 101, the estimation unit 204 is configured to calculate a degradation value as a degradation estimation of the power storage element, but is not limited thereto. The estimation unit 204 may calculate a level indicating degradation of the power storage element, a lifetime of the power storage element, a replacement timing of the power storage element, and the like.

In the estimation device 101, the estimation unit 204 may estimate the deterioration of the power storage element based on the sum of the deterioration value obtained in the present embodiment and the non-energization deterioration value Qcnd calculated by the method of patent document 1.

In the above embodiment, the case where the degradation coefficient k is determined from the waveform of each fluctuation width when the center SOC is 60% has been described, but the present invention is not limited to this.

The estimation means is not limited to the case of estimating the deterioration using the deterioration coefficient output model. The degradation may be estimated using a learning model that outputs the cycle degradation amount in the acquisition period when the amplitude spectrogram (waveform component), the average SOC, the acquisition period during which the average temperature T, SOC changes, and the previous cycle degradation amount are input.

Further, although the description has been made using fourier transform as an example of frequency transform, other transforms (for example, wavelet transform, discrete cosine transform, etc.) may be used.

The foregoing embodiments are not limiting. The scope of the present invention is intended to include meanings equivalent to the scope of the claims and all modifications within the scope.

Description of the symbols

20 a control unit;

an acquisition unit (201);

202 a decomposition part;

203 a determination section;

204 an estimating unit;

23 a storage section;

231 a degradation estimation procedure;

233 a training data DB;

234 a degradation coefficient output model;

41 a current sensor;

42 a voltage sensor;

43 a temperature sensor;

53A/D conversion part;

54 history record making part;

a 55 counter;

56 a storage section;

57 a communication unit;

60 a recording medium;

101 an estimation device;

151 monitor the device.