Determination of aging of power storage systems

Abstract

The present disclosure encompasses a method for determining the aging of a battery having a variable state of charge within a cycle window forming part of a nominal operating domain. The method comprises: - (S10) obtaining a trace of the state of charge of the battery over time during a time interval, wherein the trace comprises alternating actuations of charge and discharge within the cycle window; - for a plurality of actuations within the time interval, (S20) obtaining a value indicating an electrical measurement of energy exchanged by the battery during the actuation, and (S30) estimating a state of health of the battery during the actuation based on the value; and - (S40) determining the aging of the battery based on the estimated state of health.

Description

Technical Field

[0001] The present invention relates to the field of power storage systems.

[0002] More specifically, the invention covers a method for determining the degree of ageing of an electric power storage system, a computer program intended to implement such a method, a data storage device storing such a computer program and a processing circuit comprising such a data storage device. Background Art

[0003] Just like any battery, the smallest element in an electric vehicle battery (or a battery for other purposes) is called a cell. Common voltages for lithium-ion technology cells are usually between 2.5V and 3.4V, with the average value of discharge lying around 3V. This metric is related to the power (expressed in Watts) of the energy that can be recovered. The greater the power that can be extracted from the cell, the greater the value of the average cell voltage drop. Capacitance is also an essential metric. The unit of this metric is A·h (ampere-hour); it is the amount of current that can be extracted per unit time. The energy (expressed in Wh, Watt-hour) is the product of the capacity and the voltage that can be recovered during a full or partial discharge, and therefore, it essentially determines the stand-alone operating time of a battery system, such as that of an electric vehicle (EV).

[0004] Considering the world of electric vehicles, hundreds of these battery cells are assembled in complex ways (battery packs) in order to provide the electric motor with enough power to move the vehicle (series assembly) and ensure the EV has an acceptable runtime (parallel assembly).

[0005] The performance of electrochemical energy storage systems shrinks naturally during their use (known as cycle-related aging) or even independently of their use (known as calendar aging). For EVs, the aging depends, for example, on the battery cells used, the use of the EV, the climate, the location of the garage and even the color of the vehicle, which affects its temperature.

[0006] This established fact has led to the definition of a variable called "State of Health" or SOH. It has no units and is defined as the ratio of actual power to nominal power.

[0007] Therefore, for a new system, the actual electrical energy is the same as the nominal electrical energy. The SOH is 100%. For electric vehicles, this progressive loss leads to a progressive and uncontrollable loss of operating time (moderately predictable). If the operating time becomes unacceptable to the driver (for the Renault Zoé, the SOH is 75%), a new battery is installed in the vehicle and the old battery is recycled in different forms (dismantling, parts recycling, burning of metals and refurbishment).

[0008] The recycling step considers scenarios where re-use of these systems in areas where performance degradation is not prohibitive may be advantageous; second-life onboard EVs that lag behind grid applications in terms of operating hours or stationary storage to form a combination with, for example, intermittent, unpredictable photovoltaic generation.

[0009] Such a second life is economically attractive compared to the market price of a new product at the time of purchase and over the entire second life. This reuse is also in line with the EU's goal of establishing a circular economy.

[0010] Such systems can only be used if safety, performance and lifetime are satisfactory and sufficiently predictable. Therefore, SOH measurement of batteries is a key element for secondary use. The technical difficulty is to evaluate the standards for used batteries subjected to unknown use stresses for several years. Even at the cell level, the heterogeneity of these properties increases, especially from one battery pack to another after different use conditions.

[0011] We are particularly interested in the reuse of used electric vehicle batteries. In practice, the companies that own the batteries either recycle them or sell them to third parties. However, even after selling these batteries to third parties, they remain legally responsible for them and remain so until they are destroyed or recycled. One way to get value from these used batteries is to use them to produce new storage systems while maintaining their integrity (without disassembling the individual components of the battery when it is removed from the vehicle).

[0012] Determining the state of health of a battery or a cell is therefore a hurdle. Today, there is only one method that can determine this critical property. The measurement consists of a complete discharge of the battery or cell to measure its energy or capacity. Otherwise, predictive models of aging are used.

[0013] In the case of reuse of second-life batteries in stationary systems, two approaches are prevalent.

[0014] The first approach consists in disassembling the packaging in order to assess the quality of each battery cell. This is possible when the company that originally owned the batteries recycles them itself. They are characterized and then sorted according to a large number of parameters. This solution takes a lot of time, requires means to carry out the characterization and requires a new integration (electronics, welding, casing, etc.). It is then possible to reassemble the battery cells uniformly with an acceptable quality level.

[0015] Another approach is to produce second-life batteries from used battery packs without knowledge of the pack's state of health or by means of estimates derived from models specifically designed to estimate the state of health when on board an EV. In the case of the transfer of these batteries to third-party companies, it is difficult to envisage a transfer of data and models for obvious reasons of industrial secrecy. Other prediction models provided by VE manufacturers will not correspond to the new life conditions of batteries in fixed use. For these systems that remain intact, it is not even possible to discharge the system completely. The battery management system prohibits this possibility in order to optimize the service life during use in any EV. In fact, manufacturers of electric vehicles usually limit the charging window between low and high states (for example, between 20% and 80%) to ensure a longer battery life. It is therefore impossible to discharge the battery completely.

[0016] The prediction of aging in this system is even more complex and problematic, since the new usage conditions may be very different from those encountered during electric vehicle use in terms of temperature, current and cycling window, making it risky to attempt to use second-life batteries on a large scale.

[0017] Therefore, there is a need to quickly, simply and reliably assess the degree of aging of an electrical storage system formed by one or more batteries, without having to refer to the previous life of the one or more batteries. Summary of the invention

[0018] The present disclosure aims to improve this situation.

[0019] One aspect of the invention encompasses a method for determining the degree of ageing of at least one battery used in an electricity storage system, the battery having a variable state of charge within a cycle window forming part of a nominal operating domain, wherein for each battery:

[0020] - obtaining, from at least one moment of use of the battery, a tracking of the state of charge of the battery over time during a time interval starting from said moment, wherein the tracking comprises alternating invocations of charging and discharging within a cyclic window;

[0021] - for a plurality of awakenings within a time interval, obtaining values ​​indicative of electrical measurements of energy exchanged by the battery during the awakenings and estimating the state of health of the battery during the awakenings based on said values; and

[0022] -Determine the age of the battery based on its estimated state of health.

[0023] Another aspect of the invention encompasses a computer program comprising instructions for implementing the above method when the program is executed by a processor.

[0024] Another aspect of the invention relates to a computer-readable non-volatile record carrier having recorded thereon a program for implementing the above method when executed by a processor.

[0025] like Figure 2 As shown, another aspect of the invention relates to a processing circuit comprising a processor PROC100 connected to the above-mentioned non-volatile record carrier MEM200. Such a processing circuit may for example comprise an interface for communicating with a control device COM300 of an electric power storage system.

[0026] By tracking the charge state of a battery read from the moment of use, for example during regular operation of a second-life battery, the development of its state of health over time can be assessed.

[0027] Based on this evaluation, it is possible to determine the degree of ageing of the batteries used in the power storage system and to establish a track of the state of charge without knowing any previous use of the batteries before the moment of use.

[0028] Since it is possible to implement the method of the invention during the reuse of batteries, it is not necessary to characterize these batteries or the cells that make them up after one life and before reuse.

[0029] This saves time by determining the degree of battery aging during use, making it easier to reuse large-scale scrap batteries, such as those used in electric vehicles. In particular, reuse in stationary power storage systems becomes easier.

[0030] In one embodiment, the battery is previously installed and used in an electric vehicle.

[0031] Due to the limitations imposed during design, such batteries cannot be fully charged or discharged. Therefore, as previously described, the health state of the battery is determined by reference based on tracking its full charge or discharge. Since the charge state of the battery is tracked (including alternating charging and discharging), the degree of aging of the battery (including batteries that have been previously installed in electric vehicles) can be determined, and thus the degree of aging can be determined even if the battery cannot be fully charged or discharged.

[0032] In one embodiment, estimating the health status includes:

[0033] - identifying multiple arousals corresponding to a single arousal type profile in the generated trace; and

[0034] - Estimate the battery's state of health for each identified arousal.

[0035] Therefore, estimates of battery state of health are repeatable and directly comparable.

[0036] The power storage system that automatically charges during off-peak hours is used to naturally obtain a wake-up charge corresponding to a single type profile in normal use of the system.

[0037] The call for charging the power storage system, preferably by energy collected by the solar panels, can vary according to the sunshine. The call corresponding to the type profile can be repeated or performed regularly as needed to ensure repeated or regular evaluation of the health status of the battery.

[0038] In one embodiment, the time interval is comprised between one month and one year.

[0039] At monthly intervals, a sufficiently large number of estimates of the battery's state of health can be made with acceptable statistical accuracy to determine the extent of battery aging.

[0040] However, such a determination requires:

[0041] - Extract the variability of external parameters that affect battery aging, such as seasonal temperature changes, or

[0042] - Introducing corrections to the aging estimate as a function of these external parameters, which corrections can be read out or predicted individually, for example.

[0043] Given the variability of these external conditions, a one-year interval is ideal for estimating battery aging.

[0044] In one embodiment, the alternation of inducing charging and inducing discharging forms a cycle:

[0045] - Each health state estimated during arousal is related to the number of cycles completed in the time interval preceding arousal; and

[0046] -Determining the age of the battery based on the estimated state of health includes determining an indication of the number of cycles the battery completed before a time interval.

[0047] Thus, the number of cycles that the battery has undergone since one use can be determined. Based on this number of charges, the remaining life of the battery can be predicted, for example, based on the remaining time and / or the remaining number of cycles. With this prediction, the user can, for example, pre-order a new replacement battery.

[0048] In this example:

[0049] - Each health state estimated during the arousal is associated with the date of the arousal, and

[0050] -Determining the age of the battery based on the estimated state of health includes determining an indication of calendar age of the battery.

[0051] For example, the method of determining the degree of aging of a battery may be used in various ways.

[0052] In one embodiment, the method further comprises:

[0053] - Predict the progression of the battery's state of health based on the determined degree of aging.

[0054] Estimating the age of a battery is actually used to predict how quickly the battery's state of health will subsequently degrade.

[0055] The accuracy of the predictions is particularly high when determining calendar aging and cycle-related aging.

[0056] In fact, the degree of calendar aging and cycle-related aging affects the subsequent decline in the battery's state of health at different rhythms.

[0057] For example, the prediction may be performed using, for example, a prediction model by extrapolating the current estimated state of health of the battery and taking into account the determined degree of aging.

[0058] In one embodiment, the method further comprises:

[0059] - Modifying the operating parameters of the battery according to the determined degree of aging.

[0060] For example, the energy storage system's operating mode may be planned to be selected from among several possible operating modes based on the aging of the battery or a prediction of the battery's subsequent state of health.

[0061] In fact, if the predicted remaining life of the battery is long enough, or the aging degree of the battery is below a certain threshold, the difference between the upper and lower charge limit states can be planned to increase, thereby forming a cycle window.

[0062] Increasing the size of the cycling window has the effect of maximizing the run time of the battery. In this way, the battery is used to provide optimal performance.

[0063] If, on the contrary, the predicted remaining life of the battery is less than a preset threshold, or the aging degree of the battery exceeds a preset threshold, it may be planned to reduce the difference between the upper limit state of charge and the lower limit state of charge to form a cycle window.

[0064] Reducing the size of the cycling window has the effect of extending the remaining life of the battery compared to the predicted value. Thus, when the battery reaches the end of its life, its operation may correspond to an operating mode that delays its replacement.

[0065] In one embodiment, the method further comprises:

[0066] -Determine whether the battery needs to be replaced based on the determined degree of ageing.

[0067] In practice, the prediction of the ageing of the battery or the state of health of the battery can be used to estimate the remaining interval before battery replacement or the number of remaining cycles before battery replacement. Such an estimate can alert the user or supplier of the storage system and, for example, an order for battery replacement may be automatically generated. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Other features, details and advantages will emerge from reading the following detailed description and from analyzing the accompanying drawings, in which:

[0069] Figure 1 The structure of a processing circuit in an example for implementing the proposed method is schematically shown.

[0070] Figure 2 is a program diagram of a general algorithm of a computer program in an example implementation for implementing the proposed method.

[0071] Figure 3 is an example of tracking the battery state of charge over 1000 hour intervals between 3500 and 4500 operating hours.

[0072] Figure 4 The integration of the discharge power over each discharge is shown for multiple discharges of the battery.

[0073] Figure 5 An estimation of the progression of the battery's state of health during its operation is shown in an embodiment of the invention.

[0074] Figure 6 The aging degree prediction model according to the embodiment of the present invention is shown. Figure 5 Shown is a linearization of the progression of the battery's state of health during its operation.

[0075] Figure 7 A prediction model of degradation of the power storage system in an embodiment of the present invention is shown and compared with the estimated state of health. DETAILED DESCRIPTION

[0076] A power storage system comprising one or more batteries is contemplated. Each battery is a power storage element. A battery herein may refer to any type of electrochemical storage battery. The most common type of electrochemical storage battery is the lithium-ion type. Various lithium-ion chemistries such as NMC, LMO and LFP are contemplated, as are electrochemical storage batteries without lithium ions, such as solid electrolyte batteries.

[0077] Reference now Figure 2 , which shows an example general algorithm of a computer program for implementing an example embodiment of the proposed method.

[0078] The proposed embodiments are used to determine and protect the health status of a second-life stationary storage system consisting of one or more batteries of an electric vehicle (batteries that are not removed for reuse) without requiring data and models obtained and used during the primary life.

[0079] During part of the cycle, S10 obtains operating data tracked from the energy storage system.

[0080] An example of a partial cycle is a discharge or charge within a cycle window, the cycle or window being bounded by a maximum state of charge boundary (eg, 80%) and a minimum state of charge boundary (eg, 20%).

[0081] The charge or discharge profiles may be predefined or may respond to operational needs. The energy storage system operates normally by periodically introducing a wake-up type profile (monotonous charging or discharging between minimum and maximum boundaries if possible under constant current and similar temperature conditions) for a period representing reuse and ageing (from a few months to a year).

[0082] The operating data includes the state of charge of the battery over time, and may include changes in other electrical metrics over time, characterizing the transfer of electrical energy from or to the energy storage system, or more specifically, the transfer of electrical energy from or to the batteries of the energy storage system.

[0083] Such electrical quantities may be voltage, current, electrical power, impedance, etc.

[0084] To obtain good accuracy, the collection period may vary. For example, to store solar energy in mainland France, it is best to collect it over the year to integrate the effects associated with seasonal variations. Otherwise, a collection period in spring or autumn is chosen. Other treatments may be performed depending on the geographical location and climate.

[0085] exist Figure 3 An example of tracking operational data for such a battery in the form of state of charge over time is shown in Figure 1. The profile is highly repeatable and follows a fixed usage called "residential". The power storage system is charged by solar panels during the day and then discharged as needed. Each charge 10 is represented by an increase in state of charge. Each discharge 20 is represented by a decrease in state of charge.

[0086] Reference now Figure 4 , which shows the integration of the discharge power at each discharge for multiple discharges of the battery. This integration is obtained by tracking the change of the state of charge over time (e.g. Figure 3 as shown).

[0087] For each arousal, the power integration during these discharges (or charges) is used to determine the energy exchanged during the arousal 30. Alternatively, the current can be integrated during each charge or discharge, and this is used to determine the capacity exchanged per cycle.

[0088] Typically, the S20 can obtain the energy capacity exchanged in each cycle and can track its development during its service life by integrating the power or current of a charge or a discharge.

[0089] During each charge or partial discharge, not all of the electrical energy contained in each cell is used. However, even if the discharge is only partial, the battery as a whole ages to a certain extent. The energy produced and the capacity can be corrected, for example, by applying proportionality laws to represent each partial cycle completed as equivalent to a portion of a full cycle.

[0090] Thus, for example, the energy E(exchange ij) exchanged in a cycle between a first charge state i and a second charge state j may be considered to be proportional to |ij| (the absolute value of the difference between i and j), where i and j are expressed as percentages.

[0091] By using the same scaling laws as above, it is possible to calculate from the capacity exchanged in the partial cycle between the first state of charge i and the second state of charge j what capacity would have been if the cycle had been completed between 100% and 0% states of charge.

[0092] When the state of health SOH of the battery is 100%, the energy or capacity exchanged between the first state of charge i and the second state of charge j in a given partial cycle is maximum.

[0093] Over time, the health of the battery degrades. By repeating the same partial cycle over time, the exchanged energy loss ΔE and capacity loss ΔC can be observed.

[0094] For a plurality of partial cycles, S30 the battery's state of health SOH 40 is estimated based on the energy or capacity exchanged in the cycle.

[0095] Reference now Figure 5 , which shows the state of health SOH 40 of a battery for multiple discharges of the battery. Here, the state of health is determined based on the energy exchanged during each discharge process.

[0096] The state of health SOH 40 may be obtained by dividing the energy (or capacity) exchanged during a partial cycle by the energy (or capacity) exchanged by a battery with 100% of the state of health during the same partial cycle.

[0097] Alternatively, the state of health SOH is obtained from the energy (or capacity) exchange during part of the cycle, the method is:

[0098] - by determining the energy (or capacity) that will be exchanged after completing the cycle, thus between 100% and 0% state of charge, and

[0099] - by dividing the (capacity) energy thus determined by the nominal energy (or capacity) exchanged during a cycle completed between 100% and 0% state of charge by a battery with a 100% state of health.

[0100] The nominal energy and capacity are usually known and provided by the battery manufacturer.

[0101] The estimated state of health over time for a plurality of wake-ups on the battery is stored, and S40 determines the degree of aging of the battery based on the stored state of health.

[0102] There are many models that can be used to describe and predict aging.

[0103] In the embodiment shown, the model is used as a known fatigue-type model. This relates the stresses to which the battery is subjected (temperature, current, state of charge) to the damage caused by these stresses, which worsens with age.

[0104] This predictive aging model takes into account the new conditions of the second-life battery.

[0105] The model assumes that the energy loss ΔE associated with the degree of calendar aging and the degree of cycle-related aging is given by the expression The equation of the model can be formulated by considering the equation The relevant straight lines linearize the data.

[0106] like Figure 6 As shown, then through As A graphical representation of the function 51, with the parameters derived graphically and superimposed on The calculated value is 50, where:

[0107] -a is the value of the slope, equal to k 循环 ;as well as

[0108] -b is the coordinate value of the origin, equal to k 日历 .

[0109] Therefore, the model is used to eliminate the effects of cycle-related aging and calendar aging on the battery health state.

[0110] Starting from the determined degree of ageing, the S50 can predict the progression of the battery's state of health.

[0111] exist Figure 7 In FIG. 6 , the dashed line shows a prediction 61 of the development of the battery state of health over time over a 30,000 hour interval, superimposed with the battery state of health measurements 60 over a 5,000 hour interval.

[0112] Thus, by relying on the services provided by a power storage system comprising one or more batteries, the development of the state of health of the power storage system can be tracked and predicted without knowing the history of the primary life of the batteries.

[0113] By having both an indication of cycle-dependent aging and an indication of calendar aging, the change in battery health status as a function of calendar aging and cycle-dependent aging of individual components can be predicted with greater accuracy both during earlier use of the battery and during planned subsequent use.

[0114] S60 Based on the predicted development of the battery's state of health, the operating parameters can then be redefined to optimize the service provided by the battery and its lifetime, for example depending on the safeguards attached to the system.

[0115] S70 can also predict necessary system maintenance, and in particular indicate necessary replacement of batteries. Being able to provide such an indication is a driving force behind the commercial development of reusing the second life of electric vehicle batteries.

Claims

1. A method for determining the degree of ageing of at least one battery used in an electric power storage system, the battery having a variable state of charge within a cycle window forming part of a nominal operating domain, wherein for each battery the method comprises: - obtaining, from at least one moment of use of the battery, a tracking of the state of charge of the battery over time during a time interval starting from said moment, wherein the tracking comprises alternately invoking charging and discharging within a cycle window; - for a plurality of awakenings within a time interval, obtaining values ​​indicative of electrical measurements of energy exchanged by the battery during the awakenings and estimating the state of health of the battery during the awakenings based on said values; as well as -determine the battery's aging based on its estimated state of health, Among them, the alternation of awakening charging and awakening discharging forms a cycle: - Each health state estimated during arousal is related to the number of cycles completed in the time interval preceding arousal; and -Determining the age of the battery based on the estimated state of health includes determining an indication of the number of cycles the battery completed before a time interval.

2. The method according to claim 1, characterized in that: - Each health state estimated during the arousal is associated with the date of the arousal, and -Determining the age of the battery based on the estimated state of health includes determining an indication of calendar age of the battery.

3. The method according to claim 1, characterized in that The battery has previously been installed in electric vehicles for use.

4. The method according to claim 1, comprising: - Predict the progression of the battery's state of health based on the determined degree of aging.

5. The method according to claim 1, comprising: - Modifying the operating parameters of the battery according to the determined degree of aging.

6. The method according to claim 1, comprising: -Determine whether the battery needs to be replaced based on the determined degree of ageing.

7. The method according to claim 1, characterized in that Time intervals include between one month and one year.

8. The method according to claim 1, characterized in that Estimated health status includes: - identifying multiple arousals corresponding to a single arousal type profile in the generated trace; and - Estimate the battery's state of health for each identified arousal.

9. A computer-readable non-volatile record carrier having recorded thereon a program for implementing the method of claim 1 when executed by a processor.

10. Processing circuit comprising a processor connected to a non-volatile record carrier as claimed in claim 9.

Images