Predicting state of health of an electrochemical device by measuring capacity fade

Abstract

The invention relates to a method for predicting an acceleration of the capacity degradation of an electrochemical device, the method comprising: - obtaining point measurement data from a function linking the voltage across the terminals of the electrochemical device to the state of charge of the electrochemical device and obtaining a measurement of the capacity (Qt) of the electrochemical device, - calculating the derivative of the function and determining the peak of the variation of the derivative due to an inflection point of the variation of the function and characterizing a representative of the anodic capacity (Qa / 2) of the electrochemical device, - estimating the width of the peak (σ) and comparing the combination of the width of the peak and the representative of the anodic capacity to the measurement of the capacity (Qt) of the electrochemical device, and - predicting an acceleration of the capacity degradation of the electrochemical device if the combination is less than the capacity of the electrochemical device.

Description

Technical Field

[0001] This invention belongs to the field of energy storage and energy storage device lifetime estimation. Background Technology

[0002] This invention relates to any type of electrochemical device and / or system for storing energy, such as:

[0003] - Rechargeable battery

[0004] - Lithium-ion batteries

[0005] - Solid-state lithium-ion batteries

[0006] - Sodium-ion batteries

[0007] - Solid-state sodium-ion batteries and / or other batteries.

[0008] These devices may be designed for use in smartphones or laptops, or in electric vehicles (EVs), or in systems (BES) that generate electricity from solar or wind power or other energy sources.

[0009] Therefore, the present invention can be applied to all devices and systems in which an electrochemical device for storing electrical energy is present.

[0010] In such devices and / or systems, it is necessary to predict the “rapid decline” in the capacity (hereinafter referred to as the inflection point) of the electrochemical device used for energy storage, particularly using voltage measurements and voltage monitoring. Reference Figure 1 The lower right corner (Part 4) shows the inflection point, which more specifically represents the acceleration of capacity decline: typically, as illustrated in the example, it involves the change in the slope of capacity decline with the number of cycles (or more generally, the duration of use). This capacity, denoted by Qt in the attached figure, represents the total (i.e., maximum) capacity of the electrochemical device. It typically corresponds to the capacity of the cathode, which is a limiting factor for new devices. Especially in the attached figure... Figure 1-4 As can be seen, the absolute value of the slope increases in a given cycle, corresponding to the aforementioned inflection point.

[0011] To date, this capacity reduction has been observed particularly in lithium-ion batteries and is usually associated with lithium metal deposition (or “lithium plating”), but it occurs more commonly in batteries that use embedded materials as their two electrodes (anode and cathode).

[0012] Rechargeable lithium-ion electrochemical storage devices are based on reversible reactions at the positive (cathode) and negative (anode) electrodes. The so-called "intercalation" reaction is central to lithium-ion battery development because it exhibits high reversibility during charge / discharge operations.

[0013] During the "intercalation" reaction, the target ion (e.g., lithium or possibly sodium) moves between the cathode and anode in an ionic state rather than a metallic state throughout the operation of the device. For example, in the case of a rechargeable lithium-ion battery, lithium ions are released from the cathode (intercalation or deintercalation reaction) and intercalated at the anode (intercalation or intercalation reaction). During discharge, the reverse reaction occurs at the cathode (intercalation) and the anode (deintercalation). This "ionic" state of reaction allows for ensuring the reversibility and proper operation of rechargeable electrochemical devices.

[0014] This reaction is reversible as long as no metal deposition occurs. In other words, if a deposition reaction occurs in the "metallic" state ("lithium plating"), it triggers an irreversible reaction. Besides reducing device performance and significantly decreasing its capacity, "lithium plating" also promotes the formation of lithium dendrites, which can create internal short circuits and lead to safety issues. This is why it is important to minimize any metal deposition as much as possible during the operation of rechargeable electrochemical storage devices.

[0015] Because the potentials of graphite and lithium metal are similar, localized lithium metal deposition mainly occurs on the graphite anode.

[0016] The embedding capacity of an electrode is defined based on the number of crystalline sites accessible to the ionic material, or, in the case of a metal alloy, the ratio between the ionic material and the host metal. Therefore, the maximum amount of ions accepted by each electrode corresponds to its theoretical capacity. When excess ions migrate toward the anode (i.e., more ions than available sites), these ions no longer insert into the structure but instead deposit as metal on the graphite surface. Therefore, rechargeable electrochemical storage devices are typically designed with excess capacity on the anode side.

[0017] When metal deposition occurs during device operation, a significant irreversible reaction takes place, leading to a rapid decrease in capacity. An inflection point is then observed on the device's aging curve, corresponding to the aforementioned inflection point. Figure 4 An asterisk is used to indicate this. Figure 4 The area shown in gray (bottom right, section 4) represents the region where there is a risk of metal plating (e.g., lithium). More generally, inflection points can also be observed along with other degradation mechanisms, such as loss of cathode capacity, loss of liquid electrolyte, or others (e.g., in devices other than lithium-ion batteries). However, inflection points caused by lithium metal deposition can also put the device in a hazardous state (especially a fire risk). Therefore, inflection points caused by metal plating are the most important events to avoid during operation.

[0018] An example of an inflection point occurring after lithium metal deposition is described below.

[0019] The voltage distribution V of a rechargeable electrochemical storage device, as a function of its state of charge (denoted by Q in the figure), is determined by the potential difference between the cathode and anode. It is a good indicator for predicting inflection points related to lithium metal deposition. A typical example of the analysis is... Figure 1-1 As shown. During charging, curve "a" represents the anode voltage distribution, and curve "c" represents the cathode voltage distribution. Curve "b" represents the voltage of the device, corresponding to curves "c" - "a". The reversible capacity of the device then corresponds to the overlapping region, denoted as Q1. In this example, it is assumed that there is no capacity loss on the cathode side.

[0020] After degradation / aging, the reversible capacity of the device is Figure 1-2 The voltage decreases to Q2. In this case, the anode voltage shifts to the right (the curve shifts) ΔQ2 and decreases (the curve narrows). Here, the reversible capacity of the anode has decreased, but it is still greater than the capacity of the device (limited by the cathode capacity), so it has no effect on the total capacity of the device.

[0021] With further aging, the reversible capacity decreases to Q3, such as Figure 1-3 As shown. In this case, the reversible capacity of the anode (which becomes the limiting electrode and now determines the total capacity of the device) is less than that of the cathode. Therefore, lithium ions are in excess compared to the available insertion sites on the anode, and the anode potential reaches 0V, leading to a significant risk of metal deposition.

[0022] Figure 1-4 The typical trend of capacity Qt changing with charge / discharge cycles (or operating time) of the electrochemical device described above is shown. Between Q1 and Q2, the slope of capacity loss is constant because it depends on the value taken by ΔQ2. Then, the slope of capacity loss changes, forming an inflection point: this is the inflection point. Therefore, in this case, the capacity loss now depends on the anode capacity loss. This is a typical example showing the appearance of the inflection point caused by anode degradation.

[0023] Since the anode is usually designed to have a larger capacity than the cathode to avoid metal plating, it is difficult, if not impossible, to directly measure the anode capacity based on changes in the device voltage in practice.

[0024] To estimate anode capacity, derivative analysis (dV / dQ) of voltage versus capacity (dQ) is typically used. (Reference) Figure 2 This is the variation of dV / dQ drawn with bold black lines (in each) Figures 1 to 3 (On the right side). For example, anode materials (typically graphite) exhibit a voltage distribution consisting of plateaus during lithium-ion insertion and extraction. Figure 2As shown, the most pronounced plateau is observed when the anode capacity is half Qa / 2. Charging is assumed to begin from complete delithiation of the anode (when there are almost no lithium ions in the graphite structure). Therefore, the capacity measured between the start of charging and the highest peak of dV / dQ is Qa / 2. The reversible capacity of the anode Qa can then be estimated based on the position of this peak, i.e., Qa = 2 x Qa / 2, as shown. Figure 2-1 As shown.

[0025] In such Figure 2-2 Following the degradation shown, dV / dQ analysis can be used to estimate the loss of anode capacity. Then, the appearance of the inflection point can be predicted using dV / dQ analysis and the estimated anode capacity after degradation. By plotting the anode reversible capacity as a function of charge / discharge cycles or days of use, the inflection point can be predicted by extrapolating the trend of the Qa capacity and finding its intersection with the device capacity. In other words, the inflection point corresponds to the point where the total device capacity Qt equals the anode capacity Qa (the limiting electrode is no longer the cathode but becomes the anode).

[0026] This estimation of the inflection point appearance is based on the capacity between the two peaks to obtain the anode value at half the capacity Qa / 2. However, experience shows that in practice, the inflection point may occur before this technical prediction. In other words, the inflection point may occur when the reversible capacity of the anode Qa is still greater than the device capacity Qt. Therefore, additional considerations must be made to more accurately predict the inflection point for practical use. Summary of the Invention

[0027] This invention improves upon this situation.

[0028] To this end, the present invention proposes a method for predicting the acceleration of capacity degradation in electrochemical devices (detected by observing the slope or inflection point change of the aforementioned inflection point type during capacity evolution), the method comprising:

[0029] - Point measurement data are obtained from a function that links the voltage (V) across the terminals of the electrochemical device to the state of charge (Q) of the electrochemical device, and the (total) capacity of the electrochemical device is measured.

[0030] - Calculate the derivative of the function, determine the peak value of the derivative change due to the inflection point of the function change, and characterize a representative quantity (Qa / 2) of the anode capacity of the electrochemical device.

[0031] - Estimate the peak width (σ) and compare the combination of peak width and anode capacity (a representative value) with a measurement of the electrochemical device's capacity (Qt), and

[0032] - If the combined capacity is less than that of the electrochemical device, then an acceleration of capacity degradation of the electrochemical device is predicted.

[0033] Therefore, this invention proposes to predict the aforementioned inflection point based on the position and broadening of the peak of the derivative function (such as dV / dQ) curve, where the broadening represents the degree of non-uniformity within the electrochemical device. By analyzing this non-uniformity, the appearance of the inflection point can be predicted (and for various applications).

[0034] In one embodiment, the function (dV / dQ) used to calculate the derivative is the change in voltage (V) as a function of the state of charge (Q) of the electrochemical device, and the representative quantity of the anode capacity is half of the anode capacity of the electrochemical device (Qa / 2), given by the abscissa of the aforementioned peak.

[0035] Of course, one can also find the singularity given by the derivative dQ / dV of the reciprocal function Q(V) (instead of the choice of the function V(Q) mentioned above).

[0036] In one embodiment, the peak width (σ) is estimated by measuring the width of the Gaussian by adjusting (or "fitting") the Gaussian to the peak.

[0037] The width σ of a Gaussian can usually be given by the cardinality of the Gaussian (which intersects the x-axis), or estimated, for example, from its width at mid-height, or in another way.

[0038] Of course, there are many variations of Gaussian fitting (Bézier curve fitting or others).

[0039] In one embodiment, the above combination is given by 2*(Qa / 2)-N.σ-∈, where:

[0040] -(Qa / 2) is a value corresponding to half the anode capacity of the electrochemical device.

[0041] -N is a natural number greater than or equal to 1.

[0042] -σ corresponds to the estimate of the peak width, and

[0043] -∈ is a selected safety threshold (e.g., equal to a percentage of the total capacity Qt of the electrochemical device, such as 10% of Qt).

[0044] For example, an integer N (e.g., N=3) can be selected based on the type of electrochemical device being used.

[0045] In one embodiment, if the above combination is less than the capacity of the electrochemical device, accelerated physical degradation of the electrochemical device is further predicted, characterized by the formation of metal deposits at the anode of the electrochemical device.

[0046] For example, an electrochemical device may include at least one lithium battery, and the deposit formed at the anode is a deposit of metallic lithium.

[0047] In the case of sodium-ion batteries, sodium deposits may also form on the anode.

[0048] This accelerated capacity degradation typically occurs in anode devices containing graphite (pure or alloy).

[0049] This situation indicates that electrochemical devices are on the verge of extinction.

[0050] This kind of extinction can be achieved, especially when the total capacity Qt equals 2*(Qa / 2)-N.σ. Therefore, it can be understood that the above term ∈ is used to provide sufficient advance warning to the user.

[0051] Furthermore, in one embodiment, the method further includes generating a warning signal if the combination described above is less than the capacity of the electrochemical device, particularly if the margin represented by a safety threshold ∈ is beginning to be reached.

[0052] In particular, a warning signal (or a second warning signal) may indicate that the electrochemical device is nearing the end of its service life (especially if the margin is exceeded).

[0053] In one embodiment, the prediction of accelerated capacity degradation of the electrochemical device can be confirmed by increasing the measured volume of the electrochemical device to exceed a threshold (which may be predefined or relative) (as shown in some embodiments below).

[0054] The present invention also relates to an apparatus for predicting the acceleration of capacity degradation of an electrochemical device, the apparatus comprising at least one processing circuit connected to the electrochemical device to implement the method described above.

[0055] refer to Figure 8 The processing circuitry of the prediction device DIS may include:

[0056] - The signal input interface IN of the electrochemical device BATT during charging or discharging. These signals are the voltage V (in volts) and the charge / discharge value Q (in ampere-hours) applied across the terminals of the electrochemical device BATT.

[0057] - A memory MEM capable of temporarily storing at least voltage and charge / discharge values, as well as instruction data from a computer program used to implement the above methods.

[0058] - The processor PROC, which is capable of cooperating with the memory MEM, particularly of reading instructions stored in memory, in order to specifically execute the calculation and comparison steps of the above-mentioned methods, and

[0059] - Output interface OUT, which works with processor PROC to provide a possible warning signal ALERT, which will be played by the human-machine interface (displayed on the screen or played as an audio signal).

[0060] The processing circuit may also include a circuit breaker. Figure 8 (not shown in the figure) is used to isolate the electrochemical device and stop its operation (charging or discharging) when the total capacity Qt is equal to 2*(Qa / 2)-N.σ (e.g., when the margin ∈ is exceeded).

[0061] The present invention also relates to a computer program comprising instructions for implementing the methods described above, which, when executed by a processor of a processing circuit (e.g., Figure 8 (The type shown).

[0062] According to another aspect, a non-transitory computer-readable storage medium is provided for storing such programs. Attached Figure Description

[0063] Other features, details, and advantages will become apparent after reading the following detailed description and analyzing the accompanying drawings, among which:

[0064] Figure 1

[0065] Figure 1 The principle of the inflection point appearance is shown, for example, from the change in the device capacity slope caused by the decrease in anode capacity.

[0066] Figure 2

[0067] Figure 2 The inflection point is predicted by analyzing the derivative dV / dQ.

[0068] Figure 3

[0069] Figure 3 The non-uniform capacity loss leading to the early inflection point is shown.

[0070] Figure 4

[0071] Figure 4 The inflection point is predicted by analyzing the derivative dV / dQ and using the Gaussian distribution σ(σ1, σ2, σ3), which characterizes the degradation of the electrochemical device (and increases with degradation).

[0072] Figure 5

[0073] Figure 5 The degradation of the device health state (SoH) is shown, with capacity loss accelerating with the number of cycles. After 4000 cycles, starting from the reference state (where SoH is fixed at 100%), there are two degradation states ("Example 1" corresponds to the point before the inflection point, and "Example 2" corresponds to the point after the inflection point).

[0074] Figure 6

[0075] Figure 6 The total capacity (Qt) and anode capacity (Qa) of the electrochemical device before and after aging are shown. The vertical bars correspond to the capacity distribution σ estimated by Gaussian fitting of the dV / dQ curve at point Qa / 2.

[0076] Figure 7

[0077] Figure 7 The dV / dQ peaks are shown relative to the reference state (top of the attached figure), for a more degraded state of the same electrochemical device (attached figure). Figure 7 At the bottom, the peak widens (by about 50%), thus creating the width of the distribution σ.

[0078] Figure 8

[0079] Figure 8 An example of a prediction device as described above is given schematically. Detailed Implementation

[0080] We first refer to Figure 3 To explain the non-uniformity of the device, Figure 3 The anode voltage and dV / dQ curves before and after the capacity degradation are highlighted.

[0081] Under conditions of uniform capacity loss, the anode degrades uniformly across its entire surface. In this case, the voltage distribution of the electrode is the same at all points in the device. Therefore, the dV / dQ curve does not broaden after device aging, as... Figure 3-2 As shown.

[0082] On the other hand, if the capacity loss at the anode is non-uniform (depending on the region, with more or less significant degradation), the electrode voltage will differ at different locations. Some regions will retain high capacity, while others will lose more capacity. This situation is as follows: Figure 3-3 As shown. Both cases exhibit slope changes, but if the degradation is non-uniform, it is best to consider the variability of the anode voltage to predict the arrival of the inflection point with good reliability. This type of non-uniform degradation is particularly applicable to large-capacity devices due to the large reaction surface and / or the drastic variations in thermal conditions within the electrochemical device, but it can also occur for smaller devices. The corresponding dV / dQ curves show that the peak broadens due to the significant variation in voltage across the anode's entire surface. Figure 7 ).

[0083] Then, the present invention proposes to predict the appearance of the inflection point by using the degree of non-uniformity present within the device, such as... Figure 4As shown. The degree of non-uniformity can be estimated by adjusting the peak value corresponding to half the anode capacity, Qa / 2. The peak value can be extrapolated using a Gaussian curve.

[0084] Figure 4-4 An implementation of this embodiment is illustrated to predict the risk of inflection points, such as those associated with metal deposition at the anode. In the prior art, inflection points are only determined by… Figure 4-4 The Qa / 2 value (the position of the corresponding peak on the dV / dQ curve), represented by the gray dashed star, is estimated. In this invention, the distribution of Gaussian peaks (the position and width of the corresponding peak on the dV / dQ curve) can be used to detect the risk of metal deposition, such as... Figure 4-4 The gray cone-shaped region is shown in the diagram. The inflection point risk estimated in this way corresponds to the intersection of the extrapolation of line Q1-Q2 and the gray region, as determined by... Figure 4-4 The black asterisk (represented by a black asterisk) indicates the inflection point. Therefore, it can be understood that the inflection point prediction within the context of this invention (represented by a black asterisk) may be earlier than the predictions of the prior art (represented by a gray asterisk).

[0085] In a specific example, a set of 200Wh lithium-ion batteries was used as a sample. The cathode consisted of layered oxide and a graphite anode. Cycling tests were conducted at 10°C and 1C (C is the current applied according to the battery's nominal capacity). Battery capacity was measured before and after cycling at 25°C, C / 2, and C / 25. The capacity curve at C / 2 as a function of the complete equivalent cycle number is shown below. Figure 5 As shown. The dV / dQ curves (at 25°C and C / 25) were also calculated and fitted using Gaussian curves before (reference example) and after cycling, more precisely, before the inflection point (first working example) and after the inflection point (second working example). The distribution σ fitted by the Gaussian curve was used to estimate the anode capacity margin (excess capacity on the anode compared to the cathode capacity) before the risk of lithium metal deposition.

[0086] The first battery was removed before cycling, and the second battery was removed after cycling and placed in a glove box under argon atmosphere.

[0087] In the reference example, the initial battery capacity before cycling is defined as 100% of the healthy state of charge (SoH). The anode capacity estimated based on the Qa / 2 value is 120.6% compared to the battery capacity measured at C / 25. The anode capacity distribution estimated based on Gaussian fitting is 4.8%. Therefore, the anode capacity margin associated with the risk of lithium deposition is estimated at 15.8%. Thus, the anode capacity margin is sufficient, and the risk of lithium metal deposition and accelerated device degradation is very limited.

[0088] Furthermore, in the case of the reference sample, no metal deposits were found on the anode surface after the battery was removed.

[0089] In the first working example, after 4182 fully equivalent cycles, the measured SoH at C / 2 was 69.6%, and the measured SoH at C / 25 was 79.4%. The anode capacity estimated based on the Qa / 2 value is 99.0% compared to the battery capacity at C / 25 before cycling. The anode capacity distribution estimated based on Gaussian fitting is 19.2%. Therefore, the capacity margin relative to the risk of lithium metal deposition is estimated to be 0.4%.

[0090] Based on this small margin, it can already be predicted that degradation may accelerate and an inflection point may be observed due to the high risk of lithium metal deposition. Therefore, the inflection point can be predicted even before the actual acceleration of battery capacity loss.

[0091] Battery degradation due to lithium metal deposition is typically accompanied by changes in battery volume. Here, the average thickness of the battery is typically 122.9% (an increase in volume of approximately 23%) compared to the thickness of a reference battery.

[0092] In the second working example, after 4206 fully equivalent cycles, the measured SoH at C / 2 was 65.9%, and the measured SoH at C / 25 was 78.0%. The anode capacity estimated based on the Qa / 2 value was 92.8% compared to the battery capacity measured at C / 25 before cycling. The capacity distribution estimated based on Gaussian fitting was 16.6%. Therefore, the capacity margin associated with the risk of lithium deposition was estimated at -1.8%, indicating a high risk that lithium metal deposition has already occurred.

[0093] Compared to the reference cell, the average thickness of the cell is 134.6% (an increase in volume of approximately 35%).

[0094] During the disassembly of the battery, powdery deposits were observed on the anode surface. Upon reaction with ethanol, the deposited powder was identified as metallic lithium, confirming that a lithium plating mechanism had occurred.

[0095] These results demonstrate that lithium deposition can be predicted, and therefore, in the first working example, the inflection point can typically be predicted before it actually occurs.

[0096] This invention considers the variation in anode capacity by analyzing the distribution of the dV / dQ curve at point Qa / 2. The width of the corresponding peak (e.g., at the midpoint of the height) is... Figure 6The error bars represent the variability of the anode capacity within the device in the reference example (at SoH = 100%) and Examples 1 and 2. In Example 1, the lower limit of the error bar is close to the battery capacity. This indicates a high risk of future lithium metal deposition. In the second working example, the lower limit of the error bar is less than the battery capacity, indicating a very high risk of triggering the lithium metal deposition mechanism (and exceeding the inflection point). Lithium deposition observed after turning on the battery confirms that this mechanism has indeed occurred. Therefore, based on Example 1 above, the analysis within the scope of this invention is suitable for identifying lithium deposition in real-world conditions, but primarily suitable for predicting the degree of risk of lithium deposition (and thus the acceleration of capacity loss) by analyzing the anode capacity distribution under a given healthy state.

[0097] Of course, the present invention is not limited to the exemplary embodiments described above.

[0098] For example, it is not limited to observing changes in dV / dQ. The heat flux density (P) generated by the battery during charging and discharging can also be used as an indicator of the degree of non-uniformity within the device. In analyzing P, a similar method can be used, based on dP / dQ instead of dV / dQ. Inverse difference analysis, such as dQ / dV and dQ / dP, can also be applied. A similar analysis can also be performed using a combination of voltage and heat flux density, dV / dP and dP / dV.

[0099] This invention can also be implemented by applying the above procedure to changes in battery thickness. For example, appropriate indicators can be obtained by combining the evolution of anode capacity distribution and the evolution of battery thickness. By including thickness tracking, the accuracy of inflection point risk prediction can be improved.

[0100] Furthermore, the above experimental examples used lithium-ion batteries. However, the present invention is not limited to this type of battery. For example, in the case of sodium-ion batteries, a similar graphite anode is used. Therefore, the same procedure is also applicable to sodium-ion batteries.

[0101] This invention can also be applied to batteries with alloy anodes (e.g., silicon and / or tin). For alloy anodes, there are no obvious dV / dQ peaks during battery charging or discharging; therefore, it is best to use dQ / dV curves to estimate the battery capacity distribution.

[0102] The results obtained demonstrate that the analysis within the scope of this invention can effectively estimate the risk of inflection points and thereby estimate the lifespan of electrochemical devices.

[0103] In fact, the inflection point prediction in the sense of this invention can be used to assess the lifespan of electrochemical devices. Under current technology, linear functions or other mathematical functions are typically used to extrapolate the actual capacity curve (based on the number of cycles or days since commissioning) to estimate the remaining lifespan of the device. This extrapolation method does not allow for estimating the acceleration of capacity loss based on the inflection point. By estimating the degree of risk of the inflection point occurring, the actual remaining lifespan of the electrochemical device can be estimated.

[0104] Therefore, for example, this invention can provide electric vehicle users with more accurate health status estimates and battery life predictions. For instance, it can also more accurately determine the resale value of electric vehicle batteries, taking into account their next lifespan in stationary storage systems. It can also alert operators when maintenance or replacement is needed (especially before an accident risk occurs).

Claims

1. A method for predicting the acceleration of capacity degradation in an electrochemical device, the method comprising: - Point measurement data are obtained from a function that links the voltage between the terminals of the electrochemical device to the state of charge of the electrochemical device, and the capacity of the electrochemical device is measured. - Calculate the derivative of the function and determine the peak value of the change in the derivative due to the inflection point of the change in the function, where the inflection point represents a representative measure of the anode capacity of the electrochemical device. - Estimate the peak width and compare the combination of the peak width and a representative amount of anode capacity with measurements of the capacity of the electrochemical device, and - If the combined capacity is less than that of the electrochemical device, then an acceleration of capacity degradation of the electrochemical device is predicted.

2. The method according to claim 1, characterized in that, The function used to calculate the derivative is the voltage change as a function of the state of charge of the electrochemical device, and the representative quantity of the anode capacity is half of the anode capacity of the electrochemical device, given by the abscissa of the peak value.

3. The method according to claim 1, characterized in that, The peak width is estimated by fitting a Gaussian to the peak and measuring the width of the Gaussian.

4. The method according to claim 1, characterized in that, The combination is given by the following formula: 2*(Qa / 2)-N.σ-∈, in: -(Qa / 2) is a value corresponding to half the anode capacity of the electrochemical device. -N is a natural number greater than or equal to 1. -σ corresponds to the estimate of the peak width, and -∈ is the selected security threshold.

5. The method according to claim 1, characterized in that, If the combined capacity is less than that of the electrochemical device, degradation of the electrochemical device is further predicted, characterized by the formation of metal deposits at the anode of the electrochemical device.

6. The method according to claim 5, characterized in that, The electrochemical device includes at least one lithium battery and the deposit formed at the anode is a deposit of metallic lithium.

7. The method according to claim 5, characterized in that, The anode consists of graphite.

8. The method of claim 1, further comprising generating a warning signal if the combination is less than the capacity of the electrochemical device.

9. The method according to claim 8, characterized in that, The warning signal indicates that the electrochemical device is nearing the end of its lifespan.

10. The method according to claim 1, characterized in that, The prediction of accelerated capacity degradation of electrochemical devices was confirmed by the increase in the measured volume of the electrochemical devices to exceed a threshold.

11. An apparatus for predicting the acceleration of capacity degradation of an electrochemical device, comprising at least one processing circuit connected to the electrochemical device to implement the method according to any one of claims 1 to 10.

12. A computer medium memory storing instructions for a computer program, which, when executed by a processor of a processing circuit, causes the implementation of the method according to any one of claims 1 to 10.

Images