• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method for determining the capacitance C of an energy storage device, in particular an electrochemical battery, based on a calculated state of charge SOC and formed by the temporal integration of the current i charge Q of the energy storage device, characterized in that the capacity C of the energy storage device iteratively calculated is measured by current current i in each computing step for a current time k and the state of charge SOC of the energy storage device is calculated and in each calculation step, a capacity estimation is performed by the integrated current Q and the integrated current measurement error Q_off set in relation to the calculated state of charge SOC becomes. In order to be able to determine the capacity C of the energy storage device as simply and accurately as possible, it is provided that the capacity C of the energy storage device is calculated iteratively by the actual current i being measured and the state of charge SOC of the energy storage device being calculated in each calculation step for a current time k and In each calculation step, a capacity estimation is performed by setting the integrated current Q and the integrated current measurement error Q_off in relation to the calculated state of charge SOC.
    公开号:AT512745A1
    申请号:T50104/2012
    申请日:2012-03-28
    公开日:2013-10-15
    发明作者:Yingzhe Gu；Can Kurtulus；Markus Dohr；Philipp Breitegger；Bernhard Kortschak
    申请人:Avl List Gmbh；
    IPC主号:
    专利说明:

    1 56364
    The invention relates to a method for determining the capacitance C of an energy storage device, in particular an electrochemical battery, based on a calculated state of charge SOC and a charge Q of the energy storage device formed by the temporal integration of the current i.
    The capacity of a battery is usually calculated according to the following equation: c AQ Q2-Q1 ASOC SOC2 - SOCl 'where charge Q is the charge in [Ah] and SOC the charge state SOC (state-of-charge) normalized between 0 and 1, and C is the capacity C of the battery in [Ah].
    The state of charge SOC is determined by an algorithm that is not based on charge integration. The calculation of the state of charge is not the subject of the present invention. Methods for calculating the state of charge SOC can be found, for example. in the publication S. Piller, M. Perrin, A. Jossen, "Methods for state-of-charge detemination and their applications", Journal of Power Sources 96, p. 113-120, 2001. One possible method according to this would be e.g. the use of a Kalman filter for the calculation of the state of charge, whereby both current and voltage values are taken into account and the capacity in the stored model is set to a constant value.
    However, the capacity determination using the formula given has the following disadvantages: • Only two values for the charge and for the state of charge are taken into account. A punctual inaccuracy deteriorates the result accordingly. That is, it is not a method in which all measures over time have an influence on the result. • The calculation is inaccurate if there are only small changes in the state of charge SOC, or if the start and end values in the considered period 2 are coincidental. In both cases, a division takes place by a small value (SOC2-SOC1). • As a rule, measurements are therefore taken over several measurements, ideally a weighted average should be used, depending on the accuracy of the individual capacity determination. • Further, according to the calculation method from above, a deviation in the current measurement may cause large errors.
    A similar method is also presented in EP 1 220 413 Bl, wherein the capacity determination takes place only at fixed times, so for example only after a Voiladung the battery.
    Furthermore, US Pat. No. 8,041,522 B2 presents a recursive method for estimating the capacitance, working with different-brightness values, for example with the difference in the charge state between two times or with the charge conversion in one calculation step.
    The known methods have the disadvantage that they are relatively inaccurate.
    It is the object of the invention to avoid this disadvantage and to enable in the simplest possible way an accurate estimation of the capacity of an energy storage device.
    According to the invention, this is achieved by iteratively calculating the capacity C of the energy storage device by measuring the current i in each calculation step for a current time k and calculating the state of charge SOC of the energy storage device, and performing a capacity estimation in each calculation step by performing integrated current Q and the integrated current measurement error Q_off in relation to the calculated state of charge SOC is set. Preferably, the capacitance C is calculated by means of a recursive linear regression.
    In contrast to the known methods, the method presented here uses all the information present in order to establish a relationship between charge Q and state of charge SOC. It will be all 3
    Measured variables continuously recorded and taken into account in the calculation at any time.
    In contrast to US Pat. No. 8,041,522 B1, for example, integral values are used, for example with the state of charge SOC, or with the charge Q since the beginning of the measurement. The evaluation of integral quantities offers the advantage that inaccurate numerical differentiation of the signal is avoided.
    Preferably, in the method according to the invention, the following optimization problem is solved at each point in time in order to obtain a current estimate of the battery capacity: x_opt = nanx A * x-SOC , where A =] Q Q_off
    1 IC l] and x = - i_off! C SOCO SOC ... a vector with the state of charge over time in [% / 100] Q ... a vector with the integrated current i over time in [Ah] Q_off .. a vector with the effect of a deviation in the current measurement of 1A in [h] I ... a vector, all elements have the value 1 C ... a scalar for the capacity of the battery in [Ah] i_off ... Skalar for the offset in the current measurement in [A] SOCO ... is a scalar for the initial state of charge of the battery in [% / 100].
    An offset i_off in the current measurement is thus explicitly taken into account in the method presented here. This is important insofar as even a small deviation 4 in the current i over time can lead to large errors. The offset i_off in the current measurement can be set to a constant value or limited to a maximum or minimum value.
    Importantly, it is not necessary to store the above vectors (Q, Q_off, SOQ) because the optimization problem can be solved iteratively and very efficiently over time.As an iterative method, Recursive Least Squares For this, the optimization problem can be reshaped and the calculation of x_opt can be made as follows: 1) δ »= βn600 2) e_ <ift = e_ <#t -. + 1 · </ 3600 &amp; 1 3) β- <#. | ft Q_off, l] 1
    Qt 4) AkTbk = AkJbk_i + Q_offk SOC, 5)
    The method uses the measured current / and the determined state of charge SOC for the magazine k as input variables. From the last step, the following quantities are taken: Ak-iTAk-i (a 3x3 matrix) and Ak.iT bk-i (a 3x1 vector).
    The individual process steps are as follows (see equations above): 1. The charge Q is recalculated for the time k by adding the current current / multiplied by the time step dt to the charge in the previous calculation step. 2. Similar to step 1, the deviation of the charge Q_off caused by the notional current of 1A in the current measurement is recalculated. 3. The system matrix ATA is updated for the calculation step k with the new values Q and Q_off. 5 4. The vector ATb is updated for the current calculation step k with the new values Q, Q_off and SOC. The calculation of the state of charge SOC is not part of the present invention. 5. For the current time step, an optimal solution of the optimization problem is calculated. The currently best estimate of the capacity can be read from the solution vector x_opt. The capacity is equal to the inverse of the first entry in x_opt.
    In order to achieve an improvement in the accuracy, it may further be provided that a capacity determination is only carried out if predefined boundary conditions are met. Thus, it is advantageous if the capacity determination is carried out only after a defined minimum measurement duration tmin /, wherein preferably the minimum measurement duration tmjn is at least 1000 seconds. This avoids numerical problems at the beginning of the measurement.
    An improvement in the quality of the capacity determination can be further achieved if the optimization problem is based on a truncated singular value decomposition (TSVD). is regularized. As a further condition it can be provided that the unknown offset in the current measurement, preferably between - 200mA and + 200mA, is limited.
    The recursive calculation can possibly also be extended by means of a forgetting factor and a weighting of the measured values.
    The measured current i can also be corrected before calculating the battery charge Q by a current ikorr caused by secondary reactions (for example gassing of lead-acid batteries or NiMH batteries). This allows an even more accurate capacity determination can be performed.
    The invention will be explained in more detail below with reference to an exemplary embodiment. FIG. 1c shows the calculated state of charge SOC of an energy storage device plotted against time t, FIG. 1b plots charge Q over time t, FIG. 1c shows the integrated current offset i_off over time t, FIG. State of charge SOCO of the 6
    Energy storage device and Fig. 2 the determined by the inventive method capacity C of the energy storage device, Plotted over the time t.
    In the calculation presented here, at each time point k, the following optimization problem is solved to obtain a current estimate of the battery capacity: x _opt = hm »^ * x-SOC | With
    A = ß Q_off 1] 1 IC x
    -i.offiC
    SOCO
    The vector x_opt contains the solution of the optimization problem and consists of the determined capacity C of the battery, the estimated offset in the current measurement (i_off) and the initial state of charge of the battery (SOCO).
    The quantities are defined as follows: • SOC ... Vector with the state of charge over time in [% / 100] (Figure la) • Q ... Vector with the integrated current over time in [Ah] (Figure lb) • Q_off ... vector with the effect of a deviation in the current measurement of 1A (in [h], Fig. Lc) •! ... Vector, all elements have the value 1 (Fig. Id) • C ... Scalar for the capacity of the battery in [Ah] • i_off ... Scalar for the offset in the current measurement in [A] • SOCO. .. scalar for the initial state of charge of the battery in [% / 100] 7
    Simplified, one can imagine the solution as a graphical problem, as will be explained with reference to FIG. 1. For this purpose one tries to represent the curve SOC in FIG. 1a as well as a linear combination of the curves Q, i_off and 1 in FIGS. 1b, 1c and 1d. The coefficients of the linear combination for the best match in the sense of a quadratic deviation are then exactly the entries in the vector x_opt. This method is also known as linear regression.
    In the first 1000 seconds of the example given, very little information is available. This manifests itself in numerical problems in solving the optimization problem, that is, the system matrix ATA is poorly conditioned. For this reason, it is necessary to "regularize" the problem. For this purpose, the following methods are proposed (details can be found in C.R. Vogel, Computational Methods for Inverse Problems, 2002): • Capacity determination is only performed if enough information is available (Figure 2, curve 2). • There is a regularization of the problem using a "Truncated Singular Value Decomposition". (TSVD, Fig. 2, curve 3). • The unknown offset i_off is limited in the current measurement, e.g. between -200mA and 200mA (Figure 2, curve 4) because higher deviations are unrealistic.
    As can be seen from Fig. 2, the results of the various methods differ only in the first minutes of the calculation. After a certain time, a good estimate of the capacity C of the energy storage device is obtained by all methods.
    Furthermore, the recursive calculation can also be extended by means of the forgetting factor and with a weighting of the measured values.
    The measured current i can also be corrected before calculating the charge Q by a current ikorr caused by secondary reactions (for example, gassing of lead-acid batteries or NiMH batteries). This allows an even more accurate capacity determination can be performed.
    权利要求:
    Claims (16)
    [1]
    1. A method for determining the capacitance C of an energy storage device, in particular an electrochemical battery, based on a calculated state of charge SOC and formed by the temporal integration of the current i charge Q of the energy storage device, characterized in that the capacity C of the energy storage device iteratively calculated is measured by current current i in each computing step for a current time k and the state of charge SOC of the energy storage device is calculated and in each calculation step, a capacity estimate is performed by the integrated current Q and the integrated current measurement error Q_off set in relation to the calculated state of charge SOC becomes.
    [2]
    2. The method according to claim 1, characterized in that the capacitance C is calculated by means of a recursive linear regression.
    [3]
    3. The method according to claim 1 or 2, characterized in that the capacity C of the energy storage device from the solution of the following optimization problem is determined: jc _opt = miny | i4 * x- SOC , MC with A = Q Q_off l] and x = i_offtC SOCQ, where SOC ... is a vector with the state of charge over time in [% / 100] Q ... a vector with the integrated current i over time in [Ah] Q_off ... a vector with the effect of a deviation in the current measurement of 1A in [h] Ϊ ... a vector, all elements have the value 1 9 C ... a scalar for the capacity of the battery in [Ah] i_off ... a scalar for the offset in the current measurement in [A] SOCO ... is a scalar for the initial state of charge of the battery in [% / 100].
    [4]
    Method according to claim 3, characterized in that the charge Qk for the current time k, is calculated by adding the current i multiplied by the time step dt to the charge Qk.i of the previous step, according to the following equation: Qt = Q ^ + if * l 3600.
    [5]
    5. The method according to claim 4, characterized in that the deviation of the charge Q_off, caused by a defined notional current of preferably 1A, in the current measurement, according to the following equation: Q-Offk = Q_offk_i + dt / 36Q0.
    [6]
    6. The method according to claim 5, characterized in that the system matrix ATA for the current computing step k is updated with the new values for the charge Q and the deviation of the charge Q_off :, according to the following equation: A 1 A - A TA + Ak ~ Ak-1 Ak l + 'Qt Q.- ° St 1 Eat. A method according to claim 6, characterized in that the vector ATb for the current calculation step k is updated with the new values for the charge Q, the deviation of the charge Q_off and the state of charge SOC according to the following equation: 10 Qt i + Q-off * -soct.
    [7]
    8. The method according to claim 7, characterized in that an optimal solution of the optimization problem is calculated for the current time, according to the following equation:


    [8]
    9. The method according to any one of claims 1 to 8, characterized in that a capacity determination is performed only if predefined boundary conditions are met.
    [9]
    10. The method according to claim 9, characterized in that the capacity determination is performed only after a defined minimum measurement duration tmm, wherein preferably the minimum measurement time tmin is at least 1000 seconds.
    [10]
    11. The method according to any one of claims 1 to 10, characterized in that the optimization problem is regularized by means of a shortened singular value decomposition.
    [11]
    12. The method according to any one of claims 1 to 11, characterized in that the unknown offset i_off in the current measurement, preferably between - 200mA and + 200mA, is limited.
    [12]
    13. Method according to claim 1, characterized in that the measured values of the current i and / or the determined values of the charge SOC are weighted by means of at least one forgetting factor and / or a weighting factor, wherein preferably more recent measured values are weighted higher than older.
    [13]
    14. The method according to any one of claims 1 to 13, characterized in that the current i before the integration by a correction current ii «, n · corrected 11, wherein the correction current is determined due to secondary reactions in the energy storage device.
    [14]
    15. The method according to any one of claims 1 to 14, characterized in that the offset i_off is set in the current measurement to a constant value or limited to a maximum or minimum value. 2012 03 28 Fu 1 0.5 °, Q [10 0 -10, «· *« ♦♦ · «« · · * + »•« * • * «· · • C [-] -1- -1- -1 - 1 _ 1 _ 1 _ 1 _ 1 _ 1 _ 1 _ 1 _ 1 _ 1. 1 1 1 1 1 _, _ L, ___ -o ^ =. 1 l --- i 1 l) 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Ah] [I ~ I | 1 --4-- 1 1 -1 --- 1 --- | - * 7 1 Τ Ι -4-- 1 --- 1 III 1 1 I-1 ------- 1 .___ J_____J____I_1________1____1______1____l___1_1 3 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 3ff [A] -r T Γ ™ "1 T" I

    3500

    4000 4500 5000 1 [-] i-A i 500 1000 1500 2000 2500 3000 3500 4000 4500 Fiq-1 5000 t [s] C [Ah]

    I handed in later 3 1 ................. 56364vlp. :: :: ... * :. Aktenz .; a 50104/2012 *. ♦ * * ♦ ** .. «* ... * * ♦ · *" Assay: G 01 R (naui) PATENT CLAIMS 1. Method for determining the capacitance C of an energy storage device, in particular an electrochemical device Battery based on a calculated state of charge SOC and a charge Q formed by the temporal integration of the current i of the energy storage device, characterized in that the capacity C of the energy storage device is iteratively calculated by the actual current i measured in each calculation step for a current time k and the state of charge SOC of the energy storage device is calculated, and in each calculation step, a capacity estimation is performed by setting the integrated current Q and an integrated current measurement error Q_off in relation to the calculated state of charge SOC. 2. The method according to claim 1, characterized in that the capacitance C is calculated by means of a recursive linear regression. 3. The method according to claim 1 or 2, characterized in that the capacity C of the energy storage device is determined from the solution of the following optimization problem: x_opt = min * | A * x-SOC , where A = Q Q_off MC l] and x = ~ i _off / C SOC 0 SOC ... a vector with the state of charge over time in [% / 100] Q ... a vector with the integrated current i over time in [Ah] Q_off ... a vector with the effect of a deviation in the current measurement of lAin-tH-CAhJ Ϊ ... a vector, all elements have the value 1 REPLACED i 3 «* φ · • φ • φ • ♦ ·· φ a scalar for the capacitance of the ' Bötterie ti * i [* Ah] i_off ... a scalar for the offset in the current measurement in [A] SOCO ... is a scalar for the initial state of charge of the battery in [% / 100]. 4. The method according to claim 3, characterized in that the charge Qk for the current time k, is calculated by adding the current i multiplied by the time step dt to the charge Qk_i of the previous step, according to the following equation: Qk = Qk - + h <# / 3600. 5. The method according to claim 3 or 4, characterized in that the deviation of the charge Q_off / caused by a defined notional current of preferably 1A, in the current measurement, according to the following equation: Q_offk = Q_ofk_] + l-dt / 360Q. Method according to claims 4 and 5, characterized in that the system matrix ATA for the current calculation step k is updated with the new values for the charge Q and the deviation of the charge Q_off, according to the following equation: 4t 4t - 4t-1 -4m + Qu Q-offk 1 Iß, β_ <# * 1] · 7. A method according to claim 6, characterized in that the vector ATb for the current calculation step k has the new values for the charge Q, the deviation of the charge Q_off and the state of charge SOC is updated according to the following equation: FOLLOWED I 3 I 3 • · · φ φ φ φ φ φ φ φφφ φφφ φ φ φ φ φ

    8. A method according to claim 7, characterized in that for the current time an optimal solution of the optimization problem is calculated, according to the following equation:

    9. The method according to any one of claims 1 to 8, characterized in that a capacity determination is performed only if predefined boundary conditions are met. 10. The method according to claim 9, characterized in that the capacity determination is performed only after a defined minimum measurement duration tmm, wherein preferably the minimum measurement time tmin is at least 1000 seconds. 11. The method according to any one of claims 3 to 10, characterized in that the optimization problem is regularized by means of a shortened singular value decomposition. 12. The method according to any one of claims 3 to 11, characterized in that the unknown offset i_off in the current measurement, preferably between - 200mA and + 200mA, is limited. 13. Method according to claim 1, characterized in that the measured values of the current i and / or the determined values of the charge SOC are weighted by means of at least one forgetting factor and / or a weighting factor, wherein preferably more recent measured values are weighted higher than older.
    [15]
    14. The method according to any one of claims 1 to 13, characterized in that the current i before integration by a correction current ikorr REPLACED | 4% is corrected, whereby the correction stream XartJPid is determined for secondary reactions in the energy storage device.
    [16]
    15.A method according to any one of claims 3 to 14, characterized in that the offset i_off is set in the current measurement to a constant value or limited to a maximum or minimum value. Patents rws! ': 2013 07 02 Fu / St Dipl.-Ing, Mag, Micha ^ i Bsbeiuk A-1150 Vienna, WürlHhllfar @ 0rt®i-39 / i7 Tel .: f-m 1) 832 ß 33-0 Imf &gt; {Κ ζ &gt; '



    SUBSEQUENT
    类似技术:
    公开号 | 公开日 | 专利标题
    AT512745A1|2013-10-15|Method for determining the capacitance C of an energy storage device
    DE102006018208B4|2017-11-23|A method and apparatus for detecting a charged state of a secondary battery based on a neural network calculation
    DE102013216200B4|2017-03-09|Online battery capacity estimate
    AT513189B1|2014-11-15|Method for determining a control-technical observer for the SoC
    DE112015005201T5|2017-07-27|ASSESSMENT UNIT FOR A REMAINING STORED ENERGY AMOUNT, METHOD FOR EVALUATING A REMAINING STORED ENERGY AMOUNT OF A MEMORY BATTERY AND A COMPUTER PROGRAM
    DE112016006166T5|2018-09-20|State estimator and state estimator
    EP3058615B1|2017-08-16|Method and device for setting a maximum depth of discharge of an energy store for a time period
    DE102012010486B4|2019-01-03|Method and device for determining the actual capacity of a battery
    DE102008041546A1|2010-03-04|Method for calculating the state of charge of a battery
    DE10328055A1|2004-08-12|State quantity and parameter estimators with several partial models for an electrical energy storage
    DE102015011745B3|2016-09-22|Charge state determination in a motor vehicle battery
    DE112018001790T5|2019-12-12|DEVICE FOR MANAGING AN ENERGY STORAGE DEVICE AND METHOD FOR MANAGING AN ENERGY STORAGE DEVICE
    DE102012010487B4|2019-01-03|Method and device for determining the actual capacity of a battery
    DE10103848A1|2002-08-01|Determining and/or judging aging of battery by combination of selected operating parameters and /or battery states using fuzzy logic
    DE102008036159A1|2009-05-07|Battery&#39;s charging condition determining method for use during driving of vehicle, involves determining value of quasi idle voltage, and determining condition of battery depending on value of idle voltage
    AT505632B1|2009-03-15|METHOD FOR DATA TRANSMISSION
    EP2120691B1|2010-12-29|Method for determining the intensity of a workout
    DE102014210603A1|2015-12-17|A method of estimating an electric capacity of a secondary battery
    DE102012214804A1|2014-02-27|Method for determining quality information of battery cell i.e. rechargeable lithium ion battery cell, for e.g. wind-power plant, involves comparing one of battery cells in response to periodic change of current with change in current
    DE102015214128A1|2017-02-02|Method and apparatus for estimating a current open circuit voltage profile of a battery
    DE202008017529U1|2009-12-17|Reference buffer solution for the rapid calibration of pH measuring devices
    DE102011017599A1|2012-10-31|Method for operating a storage device for storing electrical energy and storage device for storing electrical energy
    AT511347B1|2012-11-15|METHOD FOR DETERMINING THE AGING STATUS OF AN ACCUMULATOR WITH A PREDETERMINED CAPACITY
    DE102015004113A1|2016-10-06|Measuring signal amplifier and method for supplying power to a measuring signal amplifier
    LU101984B1|2022-02-11|Method for local determination of at least one characteristic value of a battery cell
    同族专利:
    公开号 | 公开日
    WO2013144195A1|2013-10-03|
    AT512745B1|2013-12-15|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    JP2001078365A|1999-09-09|2001-03-23|Nippon Soken Inc|Battery capacity measuring apparatus|
    EP1220413B1|1999-09-09|2009-03-04|Toyota Jidosha Kabushiki Kaisha|Apparatus for battery capacity measurement and for remaining capacity calculation|
    WO2006057468A1|2004-11-29|2006-06-01|Lg Chem, Ltd.|Method and system for battery state and parameter estimation|
    US8041522B2|2009-12-02|2011-10-18|American Electric Vehicles, Ind.|System and method for recursively estimating battery cell total capacity|WO2015063311A1|2013-11-04|2015-05-07|Avl List Gmbh|Method for operating a hybrid vehicle|
    DE102018212545A1|2018-07-27|2020-01-30|Audi Ag|Method for monitoring a state of a battery, monitoring device and motor vehicle|JP4649682B2|2008-09-02|2011-03-16|株式会社豊田中央研究所|Secondary battery state estimation device|
    JP2011106953A|2009-11-17|2011-06-02|Honda Motor Co Ltd|Method for detecting battery capacity|CN111142025A|2019-12-26|2020-05-12|珠海格力电器股份有限公司|Battery SOC estimation method and device, storage medium and electric vehicle|
    DE102020209399A1|2020-07-24|2022-01-27|Robert Bosch Gesellschaft mit beschränkter Haftung|Method for detecting electrical error states of at least one exchangeable battery pack and/or an electrical device that can be connected to the at least one exchangeable battery pack and system for carrying out the method|
    CN112240980A|2020-12-18|2021-01-19|中海储能科技有限公司|SOC detection method based on learning algorithm|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    ATA50104/2012A|AT512745B1|2012-03-28|2012-03-28|Method for determining the capacitance C of an energy storage device|ATA50104/2012A| AT512745B1|2012-03-28|2012-03-28|Method for determining the capacitance C of an energy storage device|
    PCT/EP2013/056503| WO2013144195A1|2012-03-28|2013-03-27|Method for determining the capacity of an energy storage apparatus|
    [返回顶部]