Method for initializing battery charging status - Patents.com

Abstract

The method for estimating the state of charge of a battery 2 includes a state of charge initialization phase, which includes estimating an initial state of charge value, and includes the steps of receiving stored state of charge values ​​for a sleep mode time of the battery and computing a state of charge value from measurements of the no-load voltage at the terminals of the battery at the wake-up time of the battery, the initialization phase including the step of computing an error in the measurement of the no-load voltage due to relaxation of the battery, the error being computed as a function of the current temperature of the battery and a target relaxation duration modeled for each temperature, and the initial value of the state of charge being computed as a function of the error based on a comparison between the recorded state of charge value and the computed state of charge value.

Description

[Technical field]

[0001] The present invention relates to a method for initializing the state of charge of an electrochemical energy storage device, and more particularly to an accumulator battery. The present invention is intended for use in particular, but not exclusively, in the management of traction batteries for rechargeable hybrid and electric vehicles. [Background technology]

[0002] In the field of hybrid and electric vehicles, one of the main challenges for management systems of traction batteries is to estimate the state of charge (SOC) of the battery. The state of charge is an essential parameter in the electronic management of batteries. This parameter is representative of the amount of electrical energy that an accumulator battery can deliver at a given moment and is expressed as a percentage of its nominal capacity. An accurate estimation of the state of charge helps to optimize the performance and life of the battery. Known methods for determining the state of charge of a battery during operation include integrating the intensity of the current (charging or discharging) flowing through the battery as a function of time. In addition to the reliability and robustness of the integrator circuit, the reliability of such methods depends on a state of charge initialization phase when the battery wakes up. This initialization phase actually plays a key role in ensuring the convergence of the estimation and therefore the accuracy of the estimated state of charge.

[0003] A method for determining the state of charge of a battery, including an initialization phase of the estimated state of charge of the battery when the vehicle is started, i.e. when the battery operation begins, is known from the document EP 2101183 A1. According to this known method, a first state of charge value of the battery is calculated from the open circuit voltage (OCV) of the battery when the vehicle is started, and a second value resulting from the calculation of the state of charge when the vehicle is stopped, i.e. when the battery operation stops, is stored in a memory. A comparison is then made between the first calculated state of charge value and the second stored state of charge value, and the first calculated value is selected as the initial state of charge value if the result of the comparison is below a threshold value in absolute terms, otherwise the second stored value is selected. Thus, if the first value calculated from the open circuit voltage at the start has a low reliability, the second stored value is defined as the initial state of charge value.

[0004] However, this method for estimating the initial state-of-charge value of a battery is inaccurate in certain application scenarios, especially when the battery relaxation time is very short. Conventionally, battery charging and / or discharging periods are separated by periods of relaxation, during which no current flows through the battery. After a given relaxation time, the battery reaches a steady state. Once this steady state is reached, a no-load voltage (open circuit voltage) can be measured at the battery's terminals, which corresponds to its steady-state electromotive force. Depending on the type of battery, a relaxation time of several minutes to several hours is required to reach the steady state. More specifically, the relaxation time required to reach the steady state depends on the degree of polarization of the electrodes, which in turn depends significantly on the intensity of the current flowing through the battery, as well as on the application time of said current. Moreover, any determination of the electromotive force outside the steady state affects the accuracy of the estimation of the state of charge. Because the charge state initialization method according to the above mentioned document does not take into account the battery polarization as well as the battery relaxation time, this method should only be used after a long rest period of the battery, otherwise the method will be inaccurate. Summary of the Invention

[0005] Therefore, the present invention aims to provide a method for estimating an initial state of charge value of a battery of a motor vehicle while overcoming the limitations mentioned above.

[0006] To this end, the invention relates to a method for estimating a state of charge of a battery, comprising an initialization phase of the state of charge of the battery, comprising estimating an initial value of said state of charge, said initialization phase comprising: receiving a stored state of charge value of the battery when the battery transitions to a sleep mode; calculating a state of charge value using a measurement of the no-load voltage at the terminals when the battery is waking; Including, The initialization phase includes a calculation step for calculating an error of the measurement of the no-load voltage due to relaxation of the battery, said error being calculated as a function of a current temperature of the battery and a target relaxation time of the battery modeled for each temperature; wherein an estimate of the initial value of the state of charge is calculated as a function of the calculated error based on a comparison between the stored value and the calculated value of the state of charge.

[0007] Such an estimate provides a reliable estimate of the initial state of charge value of the battery and allows the state of charge estimate to be reset during the relaxation period, even if the relaxation period is too short for the battery to reach a steady state and, as a result, relaxation of the battery is not guaranteed.

[0008] Advantageously, the calculation of the error caused by relaxation can be done by: storing map data correlating a given state of charge of the battery with a given no-load voltage for a given temperature of the battery; determining a maximum error in the measurement of the no-load voltage corresponding to an error between the battery voltage at waking for a minimum relaxation time and the battery voltage at waking for the target relaxation time; providing error calibration data describing a change in a state of charge error margin relative to a state of charge value provided by map data in response to said maximum error; and calculating the error caused by relaxation using the no-load voltage measurement when the battery is waking and the calibration data.

[0009] Advantageously, said estimate of the initial value of the state of charge is calculated depending on whether the polarization direction corresponds to charging or discharging the battery when entering the sleep mode.

[0010] Advantageously, said estimate of the initial value of the state of charge is calculated taking into account a parameter representative of the actual sleep time of the battery in relation to a target relaxation time of the battery.

[0011] Advantageously, if the polarization direction corresponds to discharging at the time of the sleep mode transition and the stored state of charge value SOCmem at that moment is equal to or greater than the calculated state of charge value SOCocv at the time of waking, then the estimation of the initial state of charge value SOCini is: SOCini = min(SOCmem,SOCocv+Err) otherwise, SOCini=(1-a).SOCocv+a.(SOCocv+Err) where the parameter a varies between 0, when the battery sleep time is equal to the battery's target relax time, and 1, when the battery sleep time is equal to the minimum time required between a battery sleep request and a battery waking request.

[0012] Advantageously, if the polarization direction corresponds to charging at the time of the sleep mode transition and the stored state of charge value SOCmem at that moment is less than or equal to the calculated state of charge value SOCocv at the time of waking, then the estimation of the initial state of charge value SOCini is: SOCini = max(SOCmem,SOCocv-Err) otherwise, SOCini=(1-a).SOCocv+a.(SOCocv-Err) where the parameter a varies between 0, when the battery sleep time is equal to the battery's target relax time, and 1, when the battery sleep time is equal to the minimum time required between a battery sleep request and a battery waking request.

[0013] In a variant, if the polarization direction corresponds to discharging at the time of the sleep mode transition and the stored state of charge value SOCmem at that moment is smaller than the calculated state of charge value SOCocv at the time of waking, the estimation of the initial state of charge value SOCini is SOCini=(SOCocv+(SOCocv+Err)) / 2 It is calculated as follows.

[0014] In a variant, if the polarization direction corresponds to discharging at the time of the sleep mode transition and the stored state of charge value SOCmem at that moment is greater than the calculated state of charge value SOCocv at the time of waking, the estimation of the initial state of charge value SOCini is SOCini=(SOCocv+(SOCocv-Err)) / 2 It is calculated as follows.

[0015] The invention also relates to a management device for a battery, in particular a traction battery of an electric vehicle, characterized in that it includes means for implementing the method described above.

[0016] The invention also relates to a rechargeable hybrid or electric motor vehicle, characterized in that it comprises a device such as described above.

[0017] Other characteristics and advantages of the invention will be set forth in the description given below, by way of non-limiting example, with reference to the attached drawings, in which: [Brief description of the drawings]

[0018] [Figure 1]1 is a schematic diagram of a management device for a battery designed to implement the method according to the invention; [Diagram 2] 2 is a flow chart diagram of a method for initializing a battery charge state in association with the device of FIG. 1. [Diagram 3] FIG. 13 illustrates an example error calibration curve showing the change in error margin in state of charge caused by relaxation in relation to the estimated state of charge value of the battery upon waking. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] Fig. 1 shows a device including means for implementing a method according to an embodiment of the invention. The device 1 includes in particular a battery management system 10 (BMS). The battery management system 10 is linked to a battery 2. The battery management system is also linked to means 11 for measuring the temperature of the battery and to means 12 for measuring the voltage at the terminals of the battery, said means being known per se. The battery management system 10 includes a non-volatile memory 13 used to store data, in particular relating to the estimated state of charge of the battery SOCmem, and also to the polarization Signe_Polarisation of the battery when it goes into a sleep mode (charge or discharge), the duration of this sleep Rest_time, the voltage Ucell at the terminals of the battery and the temperature T of the battery when it wakes up. These data make it possible in particular to determine the relaxed state of the battery at the time of waking up, as will be explained in greater detail below.

[0020] The memory 13 also stores data on the correlation between a defined state of charge SOC of a battery for a given temperature of the battery and a defined no-load voltage OCV, for example in the form of a map OCV=f(SOC). These data can be obtained from preliminary tests on such a battery under different temperature conditions. Thus, when the battery wakes up, the measured voltage at the battery terminals is the no-load voltage OCV of the battery, and no current flows through the battery, and the battery management system 10 can calculate the corresponding state of charge SOCocv of the battery, using the measured no-load voltage OCV when the battery wakes up and the temperature of the battery.

[0021] The initialization method is intended to determine a reliable initial state-of-charge value when the battery wakes up. In reality, when the relaxation of the battery cannot be guaranteed, the state-of-charge value based on the measurement of the no-load voltage when the battery wakes up is subject to an error caused by the relaxation of the battery. Thus, steps are implemented to calibrate this error caused by relaxation, as detailed below, with the aim of obtaining an order of magnitude representative of the error occurring in the no-load voltage measurement imposed on the battery when waking up.

[0022] The error depends inter alia on the duration of the relaxation period for the battery under consideration. Furthermore, relaxation curves may be drawn, for example, at different temperatures, each curve corresponding to a particular temperature. Preferably, different curves are drawn depending on whether the relaxation period follows a battery charge or discharge polarization.

[0023] These relaxation curves for different temperatures are analyzed and a target relaxation time Thresh is identified for each temperature.

[0024] Furthermore, the maximum error Errmax of the no-load voltage measurement caused by incomplete relaxation is also determined. This error is estimated considering the error between the battery voltage at waking for a minimum relaxation time corresponding to the minimum time required between the battery sleep request and the wake request for two cycles (e.g., 1 minute) and the battery voltage at waking for a target relaxation time determined for the measured temperature of the battery. The table below shows an example of relaxation considerations performed at different temperatures to determine the corresponding target relationship time and maximum error of the no-load voltage measurement taking into account the minimum sleep / wake cycle time of the battery as explained above. TIFF0007635142000001.tif41170

[0025] The maximum error in the no-load voltage measurement is then converted to a state-of-charge error using an error calculation performed on the map OCV=f(SOC) to establish a correlation between the state-of-charge SOC of the battery and the waking voltage or no-load voltage OCV of the battery for a given temperature of the battery. This conversion of the maximum error in the no-load voltage measurement to a state-of-charge error is preferably performed offline.

[0026] Figure 3 shows the result of this conversion for an example of a determined maximum error of 50 mV in the value of the waking voltage of the battery. Thus, based on the stored reality map for a given temperature OCV=f(SOC), the determined maximum error value is taken into account for each voltage value at the waking of the battery, providing an upper and lower error margin for the corresponding state of charge value. Figure 3 plots a calibration curve for the resulting error, showing the change of this error margin SOC_error of the state of charge as a percentage in relation to the state of charge value SOC provided by the map OCV=f(SOC).

[0027] Advantageously, this error calibration curve can be decomposed into several zones, preferably at least two zones. The calibration curve shown in the example of Fig. 3 has three zones, designated with the reference symbols zone 1, zone 2, zone 3, each corresponding to a predefined range of state of charge values. Thus, the state of charge value is obtained from the map OCV=f(SOC) for the battery voltage value measured at waking, and depending on the predefined range of state of charge to which this value falls, the estimated maximum value of the error margin SOC_error for the predefined state of charge range under consideration is interpreted as the error Err caused by relaxation. According to the example of Fig. 3, for the predefined range zone 1 of the state of charge, the error Err caused by relaxation is considered to be 7.5%, for the predefined range zone 2 of the state of charge, the error Err caused by relaxation is considered to be 15% and for the predefined range zone 3 of the state of charge, the error Err caused by relaxation is considered to be 5%. Thus, taking into account the condition of the battery's voltage upon waking, with respect to belonging to one of several predefined zones, obviates the need to use a maximum error margin in the calculation, and instead adapts said error margin as closely as possible to the actual error margin.

[0028] Figure 2 is a flow chart diagram of a method for initialising the state of charge of the battery, in relation to the device required for the implementation, as described in figure 1. In a first step E1, the battery management system 10 retrieves the data recorded in the memory 13, as indicated above, namely the estimated state of charge SOCmem of the battery at the time of entering the sleep mode, the polarization Signe_Polarisation of the battery at the time of entering the sleep mode (charge or discharge), the value of the voltage Ucell at the terminals of the battery at the time of waking, the temperature T of the battery at the time of waking and finally the sleep time Rest_time.

[0029] A second step E2 involves calculating the state of charge SOCocv deduced from a map OCV=f(SOC) for the temperature T of the battery using a direct measurement of the battery voltage Ucell at waking corresponding to its no-load voltage OCV.

[0030] These two steps E1 and E2 constitute an initialization phase, which is followed by a calibration phase for the errors induced by the relaxation, implemented according to the principles presented above.

[0031] First, in step E3, the sleep time Rest_time is compared with a target relaxation time Thresh resulting from the relaxation considerations of the battery under consideration for a given temperature T. If the sleep time Rest_time of the battery is equal to or greater than the target relaxation time Thresh, then in step E4 the already calculated state-of-charge value SOCocv is set as the initial state-of-charge value SOCini. Indeed, in this case, the relaxation time required for the battery to reach a steady state has been achieved or exceeded, and therefore the voltage value OCV of the battery measured at waking provides a reliable initial state-of-charge value. Furthermore, an access control step to the memory 13 can be implemented in step E3. If the memory is not accessible, the method nevertheless proceeds with step E4 and the initial state-of-charge value SOCini is set to the calculated state-of-charge value SOCocv, since no other value is available. In other words, if the battery relaxation is complete or whatever information stored in memory is not available for any reason, an initial state of charge value is determined using the calculated state of charge value SOCocv.

[0032] Conversely, if the memory is accessible and the sleep time Rest_time is shorter than the target relaxation time Thresh, the method proceeds to step E5 for calibrating the error Err caused by relaxation. According to the principles presented above, this error is determined using the error calibration curve of Figure 3 as a function of the battery voltage value Ucell at waking.

[0033] Once the error Err caused by the relaxation has been determined in step E5, the method proceeds to a calculation phase for an initial state-of-charge value SOCini. Advantageously, the polarization direction of the battery is taken into account when calculating the initial state-of-charge value. Taking the polarization direction into account helps to get closer to the actual state-of-charge value by minimizing overestimation or underestimation errors in the state of charge.

[0034] Thus, a polarization test is first performed in step E6 to determine the polarization direction at the time of sleep mode entry. This test is based on the analysis of the change in the battery voltage immediately after sleep mode entry. A voltage increase at the very first relaxation indicates that the battery discharge has been interrupted, while conversely, a voltage decrease at the very first relaxation indicates that the battery charging has been interrupted. A variable Signe_Polarisation stored in memory indicates the result of this test. Depending on the polarization direction of the battery provided in step E6 by the variable Signe_Polarisation, a comparison is performed in step E7 between the estimated state of charge SOCmem of the battery stored in memory at the time of sleep mode entry and the calculated state of charge SOCocv resulting from a direct measurement of the OCV voltage of the battery at waking. Then, depending on this comparison, a decision is made to determine an initial state of charge value with the error Err caused by the relaxation, as determined in step E5.

[0035] If the variable Signe_Polarisation saved in the memory indicates that the battery discharge (DCH) was interrupted at the time of the sleep mode transition, then if the estimated state of charge SOCmem of the battery at the time of the sleep mode transition is greater than or equal to the calculated state of charge SOCocv at the time of waking (SOCmem>=SOCocv), then an initial state of charge value SOCini is determined in step E8 as follows: SOCini = min(SOCmem,SOCocv+Err)

[0036] Conversely, in the case where battery discharge is interrupted during the transition to sleep mode, when SOCmem < SOCocv, the initial state of charge value SOCini is determined in step E9 as follows. SOCini=(1-a).SOCocv+a.(SOCocv+Err) a is a parameter that varies between 0 and 1, represents the sleep time Rest_time, and is related to the target relaxation time Thresh of the battery for a given temperature. More specifically, a = 0 when the sleep time is equal to the target relaxation time, and a = 1 when the sleep time is equal to the minimum duration of the battery's sleep / wake cycle (e.g., 1 minute).

[0037] Taking into account the actual sleep time in relation to the target relaxation time using the parameter a helps to optimize the calculation of the initial value SOCini.

[0038] In step E6, when the variable Signe_Polarisation saved in memory indicates that battery charging rather than battery discharge was interrupted during the transition to sleep mode, and when the estimated state of charge SOCmem of the battery at the time of sleep mode transition is less than or equal to the calculated state of charge SOCocv at wake-up (SOCmem <= SOCocv), the initial state of charge value SOCini is determined in step E10 as follows. SOCini=max(SOCmem,SOCocv-Err)

[0039] Conversely, in the case where battery charging is interrupted during the transition to sleep mode, when SOCmem > SOCocv, the initial state of charge value SOCini is determined in step E11 as follows. SOCini=(1-a).SOCocv+a.(SOCocv-Err) The parameter a is determined as explained above.

[0040] In a variant, in the following example, the initial state of charge value SOCini can be calculated as follows.

[0041] When the battery discharge is interrupted during the transition to the sleep mode and SOCmem < SOCocv, SOCini = SOCocv + (SOCocv + Err) / 2

[0042] When the battery charging is interrupted during the transition to the sleep mode and SOCmem > SOCocv, SOCini = SOCocv + (SOCocv - Err) / 2

[0043] The method according to the present invention, therefore, enables the state-of-charge value to approach a more physically substantial value each time the value is initialized, thereby preventing the cumulative error of the state of charge saved in the memory.

Claims

1. A method for estimating a state of charge of a battery (2), comprising an initialization phase of the state of charge of the battery including estimating an initial value of the state of charge (SOC), the initialization phase comprising: A step (E1) of receiving a state of charge value (SOCmem) of the battery stored in a memory when the battery is switched to a sleep mode; a step (E2) of calculating a state of charge value (SOCoCv) using the measured no-load voltage (Ucell) at the terminals of the battery when the battery is waking; A method comprising: The initialization phase includes a calculation step (E5) for calculating an error (Err) of the measurement of the no-load voltage due to relaxation of the battery, said error (Err) being calculated as a function of the temperature (T) of the battery at the time of waking, the duration of the sleep mode (Rest_time) and a target relaxation time (Thresh) of the battery modelled for each temperature, If the duration of the sleep mode (Rest_time) is equal to or greater than the target relaxation time (Thresh), the calculated state of charge value (SOCoCv) is set as an initial value of the state of charge (SOCini); If the duration of the sleep mode (Rest_time) is less than the target relaxation time (Thresh), the calculation of the error (Err) is performed, the calculation of the error (Err) being: storing map data correlating a given state of charge of the battery with a given no-load voltage for a given temperature of the battery; determining a maximum error (Err_max) of the measurement of the no-load voltage corresponding to the error (Err) between a battery voltage at waking for a minimum relaxation time and the battery voltage at waking for the target relaxation time (Thresh); providing error calibration data describing a change in an error margin of the state of charge (SOC_error) in relation to a state of charge value (SOC) provided by the map data as a function of the maximum error (Err_max), the error calibration data being resolved into a number of zones respectively corresponding to a number of predefined ranges of the state of charge value (SOC) provided; calculating, as the error (Err), a maximum value of the error margin (SOC_error) in a pre-charging range to which the state of charge value (SOCoCv) calculated using the measured value (Ucell) of the no-load voltage when the battery is waking, using the error calibration data; Including, the estimate of the initial value of the state of charge (SOCini) is calculated based on a comparison between the state of charge value (SOCmem) stored in the memory and the calculated state of charge value (SOCocv) depending on the calculated error (Err) and whether the polarization direction of the battery indicates that charging of the battery was interrupted or that discharging of the battery was interrupted upon entering a sleep mode based on an analysis of a change in battery voltage immediately after entering a sleep mode.

2. If the polarization direction indicates that the discharge of the battery was interrupted at the time of the sleep mode transition, if the state of charge value (SOCmem) stored in the memory at that moment is equal to or greater than the calculated state of charge value (SOCocv), the estimation of the initial value of the state of charge (SOCini) is SOCini=min(SOCmem, SOCocv+Err) It is calculated as If not, SOCini=(1-a)・SOCocv+a・(SOCocv+Err) and parameter a varies between 0 when the duration of the sleep mode of the battery (Rest_time) is equal to the target relaxation time of the battery (Thresh) and 1 when the duration of the sleep mode of the battery (Rest_time) is equal to the minimum time required between a battery sleep request and a battery waking request.

3. If the polarization direction indicates that charging of the battery was interrupted when entering a sleep mode, if the state of charge value (SOCmem) stored in the memory at that moment is less than or equal to the calculated state of charge value (SOCocv), the estimation of the initial value (SOCini) of the state of charge is SOCini=max(SOCmem, SOCocv-Err) It is calculated as If not, SOCini=(1-a)・SOCocv+a・(SOCocv-Err) and the parameter a varies between 0 when the duration of the sleep mode (Rest_time) is equal to the target relaxation time (Thresh) of the battery, and 1 when the duration of the sleep mode (Rest_time) is equal to the minimum time required between a battery sleep request and a battery waking request.

4. If the polarization direction indicates that the discharge of the battery was interrupted when entering a sleep mode, if the state of charge value (SOCmem) stored in the memory at that moment is smaller than the calculated state of charge value (SOCocv) at the time of waking, the estimation of the initial value (SOCini) of the state of charge is SOCini=(SOCocv+(SOCocv+Err)) / 2 2. The method of claim 1, wherein the calculation is performed by:

5. If the polarization direction indicates that charging of the battery was interrupted when entering a sleep mode, the estimation of the initial value of the state of charge (SOCini) is performed if the state of charge value (SOCmem) stored in the memory at that moment is greater than the calculated state of charge value (SOCocv) at the time of waking. SOCini=(SOCocv+(SOCocv-Err)) / 2 5. The method according to claim 1 or 4, characterized in that it is calculated by:

6. A management device (1) for a traction battery of an electric vehicle, characterized in that it comprises means for carrying out the method according to any one of claims 1 to 5, linked to means for measuring the temperature of the battery and to means for measuring the voltage at the terminals of the battery, and including said memory used for storing data.

7. A rechargeable hybrid or electric motor vehicle, characterized in that it includes a management device according to claim 6.

Images