METHOD FOR EVALUATING THE RELIABILITY OF CELLS IN A VEHICLE BATTERY
Patent Information
- Application Number
- FR2023003254
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2023-04-03
- Publication Date
- 2025-02-14
- Estimated Expiration
- 2043-04-03
Abstract
Description
Title of the invention: METHOD FOR EVALUATING THE RELIABILITY OF CELLS OF A VEHICLE BATTERY
[0001] One aspect of the invention relates to a method for evaluating the reliability of cells of a battery for an electric or hybrid vehicle. For the remainder of the description, the term “battery” means a simple battery or a battery pack.
[0002] The invention finds applications for all types of batteries, for example 12V, 48V or even 400V type.
[0003] For example, batteries are known which are used to supply energy to at least one electric traction motor which the vehicle comprises and to be recharged by an electrical terminal external to the vehicle and / or by the electric motor in generator operation.
[0004] In use, the battery is subjected to numerous stress factors formed for example by the operating temperature, the frequency of charging and discharging of the battery as well as the average depth of discharge of the battery.
[0005] A used battery is very expensive to replace or even to repair, inducing very low levels of unreliability, for which adequate risk management must be carried out because it is systematically associated with a guarantee of several years, for example 8 years, and / or several tens of thousands of kilometers, for example 160,000 km.
[0006] In order to verify the reliability of a battery, a method is known for evaluating the reliability of cells of a battery consisting of determining several values of damage of the battery as a function of measurements obtained during a battery damage test carried out over a period of one year and of determining several values of loss of capacity of each of the cells as a function of capacity measurements carried out periodically during the test year.
[0007] This type of test allows the reliability of a battery to be checked but is nevertheless limited because the test can be erroneous. When the reliability level is erroneous, the battery may have a lifespan that is less than the guarantee and this is detrimental to the customer and the manufacturer.
[0008] The aim of the invention is in particular to propose a method for evaluating the reliability of cells of a more robust battery.
[0009] In this context, the invention thus relates, in its broadest acceptance, to a method for evaluating the reliability of cells of a vehicle battery, the method comprising the steps, executed by calculation means, of: • Determine several battery damage values based on measurements obtained during a battery damage test carried out over a predetermined period; Determine several capacity loss values of each of the cells based on capacity measurements made during the predetermined period; Determine, by means of a Bayesian analysis of the damage values and the determined capacity loss values, • A first principal variance of a linear regression function of a variance of a measurement error on the whole battery and of a variance of a measurement noise of the cells; • A second main variance of capacity balances depending on a variance of differences in aging speed between cells and the variance of a measurement noise of the cells for a given level of damage; • A third main variance depending on the variance of a cell measurement noise; • An analysis component of at least one corrective parameter of correspondence between said damage values and said determined capacity loss values; • Determine variance values by means of a Monte Carlo analysis of said first, second, third principal variances and of said analysis component; • Determine, by means of a statistical reliability analysis, a reliability rate for each variance value determined by means of said Monte Carlo analysis.
[0010] Thus, thanks to the reliability rates determined, it is possible to confirm that the 8-year warranty can be achieved or, on the contrary, to consider that the battery is not robust enough to ensure the 8-year warranty.
[0011] In addition to the characteristics which have just been mentioned in the preceding paragraph, the method according to the invention may have one or more complementary characteristics among the following, considered individually or according to all technically possible combinations.
[0012] According to a non-limiting aspect of the invention, the first principal variance of a linear regression is defined as follows: „2 — j. 21; — ' ea^.
[0013] With: * = First principal variance of a linear regression; • 0¾ = variance of a measurement error over the entire battery; • = variance of a measurement noise on the cells; and • Nc = a number of measurable cells in the battery.
[0014] By measurable cell is meant a single cell or at least two cells in parallel sharing the same measurement sensor.
[0015] According to a non-limiting aspect of the invention, the second main variance of capacity balances is defined as follows: aZ-(J6+N]
[0016] With: • ^z = second main variance of capacity balances; * = variance of differences in aging speed between cells of the battery; • <7fj = variance of a measurement noise on the cells; and • ~ A number of damage tests.
[0017] If I perform a year of tests with monthly verification, then Nd = 12
[0018] According to a non-limiting aspect of the invention, the third main variance is defined as follows:
[0019] With: * third main variance; and • = variance of a measurement noise on the cells.
[0020] According to a non-limiting aspect of the invention, when a variance of differences in aging rate between the cells of the battery or a variance of a measurement error over the entire battery is negative, this negative variance is ignored when performing the Monte-Carlo analysis.
[0021] According to a non-limiting aspect of the invention, the first, second, third principal variances and the analysis component are determined from a linearization of the following formula: Q (D) = rDa • Ac • • Ef /
[0022] With: • Q(D) = loss of capacity of each cell; • D = damage; . r and a _ corrective parameters for correspondence between damage values and capacity loss values; • Ac = differences in aging speed between cells; • = measurement noise on the cells; and • = measurement error across the entire battery.
[0023] According to a non-limiting aspect of the invention, the linearization is a log-linearization.
[0024] According to a non-limiting aspect of the invention, each damage to the battery is determined based on measurements of temperature, state of charge, current and depth of discharge.
[0025] Another aspect of the invention relates to a computer program product comprising instructions which, when the program is executed by computing means, cause the latter to implement the steps of the method according to any one of the aforementioned aspects of the invention.
[0026] The invention and its various applications will be better understood upon reading the following description and examining the accompanying figure.
[0027] [Fig.l] illustrates, in a schematic manner, a non-limiting exemplary embodiment of calculation means arranged to execute the steps of a method for evaluating the reliability of cells of a vehicle battery according to the invention.
[0028] More particularly, [Fig.l] illustrates calculation means 1 which can be formed, for example, by one or more computers and / or by one or more servers.
[0029] [Fig.l] also illustrates an example of execution of the steps of the method 100 for evaluating the reliability of cells of a vehicle battery according to an example embodiment of the invention, these steps being executed by the calculation means 1.
[0030] The method 100 comprises a step of determining 101 several battery damage values based on measurements obtained during a battery damage test carried out over a predetermined period. The predetermined period may for example be between 6 months and 18 months, typically 12 months. Thus, it is possible to determine battery damage based on stress factors experienced by the battery. The stress factors may be formed by temperatures, states of charge, currents and depths of discharge.
[0031] It follows that each damage of the battery can be determined based on measurements of temperatures, states of charge, current and depth of discharge.
[0032] The method 100 also comprises a step of determining 102 several capacity loss values of each of the cells of the battery as a function of capacity measurements carried out during the predetermined period. Thus, it is possible to determine the capacity loss of each of the cells as a function of the damage determined in step 101.
[0033] The method 100 further comprises a step of determining 103, by means of a Bayesian analysis of the determined damage values and capacity loss values, • A first principal variance of a linear regression function of a variance of a measurement error on the whole battery and a variance of a measurement noise of the cells; • A second main variance of capacity balances depending on a variance of differences in aging speed between cells and the variance of a measurement noise of the cells for a level of damage determined ; • A third main variance <7^ function of the variance of a measurement noise of the cells; and • An analysis component r and a of two corrective parameters of correspondence between the damage values and the determined capacity loss values.
[0034] The first, second, third principal variances and the analysis component 7^, ¢7^, <7^, r and a are determined from a linearization of the following formula: Qc(D)=rDa
[0035] With: • Q (D) = loss of capacity of each cell; • D = damage; . r and a _ corrective parameters for correspondence between damage values and capacity loss values; • Ac = differences in aging speed between cells; • = measurement noise on the cells; and • E.; = measurement error on the entire battery.
[0036] If there were no measurement noise, measurement error and capacity difference between cells, the model would be as follows: — D
[0037] The linearization of the formula is a log-linearization.
[0038] The transition to log scale allows us to obtain the following formula: 1 nO (D) = Inr + aln£) + lnAc +1nHc^ + InEj
[0039] Which we note more simply: Ycd=p+a- Xd + ôc + + ed
[0040] With: • Inr = p ; . InA = ô ; • InH — p ; and • InE = e.
[0041] In a non-limiting embodiment, to determine the first main variance 7^, the second main variance <7^ and the third main variance (7^, we determine: • An average of the cells for each capacity assessment: ^d~ P + a'^dY^+fld + ' • An average of the capacity balances for each cell: Y c — p + ci • X + Ôc + f]c + ë ; • A general average: ÿ-p + ct.X + 5+ +è
[0042] In a non-limiting embodiment, to determine the first main variance, the second main variance and the third main variance , the following linear combinations are also realized • Variability of capacity balances: — Yc - T — -Ô+ î] -1] ; ct • Variability of capacity and damage balances: Rci = Y ci - Yd-Y C + =Y^ rjcd -1)d - f}c + V
[0043] According to a non-limiting embodiment, the first principal variance of a linear regression and the correction coefficients P and a are determined by a Bayesian linear regression between Ydet Xd; The variance of the linear regression is defined as follows • ? _ ;
[0044] With: * = First principal variance of a linear regression; • = variance of a measurement error across the entire battery; • = variance of a measurement noise on the cells; and • Nc = number of measurable cells in the battery.
[0045] By measurable cell is meant a single cell or at least two cells in parallel sharing the same measurement sensor.
[0046] According to a non-limiting embodiment, the second main variance <7^ of capacity balances is defined as follows: 2 — zt2 °z “ + ncI
[0047] With: • ^2 = second main variance of capacity balances; • variance of differences in aging speed between cells of the battery; • cr^ = variance of a measurement noise on the cells; and • Nd — number of damage tests.
[0048] According to a non-limiting embodiment, the third main variance a^ is defined as follows: ai = AU
[0049] With: * <7^= third main variance; and • = variance of a measurement noise on the cells.
[0050] The method 100 further comprises a step of determining 104 variance values by means of a Monte Carlo analysis of the results of the first, second, third principal variances and the analysis component.
[0051] According to a non-limiting embodiment, when a variance of differences in aging rate between the cells of the battery or a variance of an error of measurement on the whole battery is negative, this negative variance is ignored when performing the Monte Carlo analysis.
[0052] The method 100 further comprises a step of determining 105, by means of a statistical reliability analysis, a reliability rate for each variance value determined by means of the Monte Carlo analysis. According to a non-limiting embodiment, the statistical analysis is formed by a stress-strength analysis (better known under the English term stress-strength).
[0053] Thus, thanks to the reliability rate determined, it is possible to determine whether the battery has, for a proven probability, a robustness allowing it to meet the expected time and mileage guarantees.
Claims
1. Claims Method (100) for evaluating the reliability of cells of a vehicle battery, said method (100) comprising the steps, executed by calculation means (1), of: - Determine (101) several damage values of said battery based on measurements obtained during a damage test of said battery carried out over a predetermined period; - Determine (102) several capacity loss values of each of said cells based on capacity measurements carried out during said predetermined period; - Said method (100) being characterized in that it comprises the steps, executed by calculation means (1), of: - Determine (103), by means of a Bayesian analysis of said damage values and said determined capacity loss values, • A first main variance of a linear regression function of a variance of a measurement error on the whole of said battery and of a variance of a measurement noise of said cells; • A second main variance of capacity balances depending on a variance of differences in aging speed between said cells and said variance of a measurement noise of said cells for a determined level of damage; • A third main variance depending on said variance of a measurement noise of said cells; • An analysis component of at least one corrective parameter of correspondence between said damage values and said determined capacity loss values; - Determine (104) variance values by means of a Monte Carlo analysis of said first, second, third principal variances and of said analysis component; - Determine (105), by means of a statistical reliability analysis, a reliability rate for each variance value de- completed by means of said Monte Carlo analysis.
2. Method (100) according to the preceding claim, characterized in that the first main variance of a linear regression is defined as follows: 2 + ; - With: - 0^ = first main variance of a linear regression; - = variance of a measurement error on the whole of said battery; - 0% = variance of a measurement noise on the cells; and - Nc = number of measurable cells that the battery comprises.
3. Method (100) according to any one of the preceding claims, characterized in that the second main variance of capacity balances is defined as follows: ^2 0 . N - With: - = second main variance of capacity balances; - CTg = variance of differences in aging rate between the cells of the battery; - N = variance of a measurement noise on the cells; and - Nd = number of damage tests.
4. Method (100) according to any one of the preceding claims, characterized in that the third main variance is defined as follows: = 0^ - With: - 0^ third main variance; and - N = variance of a measurement noise on the cells.
5. Method (100) according to any one of the preceding claims, characterized in that when a variance of differences in aging rate between the cells of the battery or a variance of a measurement error over the whole battery is negative, this negative variance is ignored when performing Monte Carlo analysis.
6. Method (100) according to any one of the preceding claims, characterized in that the first, second, third main variances and the analysis component are determined from a linearization of the following formula: = fD" • Ac • Hcx / • - With: - Q (D) = loss of capacity of each of the cells; - D = damage; . r and a _ corrective parameters of correspondence between the damage values and the capacity loss values; - Ac = differences in aging rate between the cells; - = measurement noise on the cells; and - = measurement error on the entire battery.
7. Method (100) according to the preceding claim, characterized in that the linearization is a log-linearization.
8. Method (100) according to any one of the preceding claims, characterized in that each damage of said battery is determined as a function of measurements of temperatures, states of charge, current and depth of discharge.
9. Computer program product comprising instructions which, when the program is executed by calculation means (1), cause the latter to implement the steps of the method (100) according to any one of the preceding claims.