Method for determining state value of power battery

Abstract

The present invention relates to a method for determining a state value of a power battery of an electric vehicle, the state value characterizing the aging state of the power battery, preferably characterizing the SoH value of the power battery, wherein the power battery is charged or discharged through a test load, and at at least one time point, a corresponding output voltage and load current value pair of the power battery is acquired, wherein an ohmic internal resistance of the power battery is established based on the acquired output voltage and load current value pair, and wherein a state value of the power battery is established based on the established ohmic internal resistance. Provided is to establish at least one standardized variable characterizing the power battery, and based on the established ohmic internal resistance and at least one standardized variable, establish a standardized internal resistance of a reference value based on the standardized variable, wherein a state value of the power battery is established based on the standardized internal resistance. The present invention also relates to a diagnostic device for determining a state value of a power battery of an electric vehicle, wherein the diagnostic device has an evaluation unit which is directly or indirectly coupled to the power battery and is configured to execute the method.

Description

Field of the Invention

[0001] The present invention relates to a method for determining a state value of a power battery of an electric vehicle, the state value characterizing the aging state of the power battery, preferably the SoH value of the power battery, wherein the power battery is charged or discharged through a test load, and at least at one time point, a corresponding output voltage and a charge or discharge current value pair of the power battery are acquired, wherein an ohmic internal resistance of the power battery is established based on the acquired output voltage and charge or discharge current value pair, and wherein a state value of the power battery is established based on the established ohmic internal resistance. Background Art

[0002] In the transportation sector, electric vehicles are increasingly being used to reduce the emission of climate-damaging gases, especially CO2. This applies not only to passenger cars but also to trucks. At least some of the electrical power required to operate an electric vehicle is temporarily stored in a power battery and made available when needed. Thus, the described power battery is a rechargeable battery or accumulator.

[0003] The described electric vehicles are not necessarily only those electric vehicles whose operating power is provided solely by a power battery. Instead, the term "electric vehicle" also encompasses those vehicles in which the electric drive is combined with another drive (such as an internal combustion engine), i.e., so-called hybrid vehicles.

[0004] The term "electric vehicle" also encompasses not only those vehicles in which the power battery is charged by an external power source, but also those vehicles in which at least some of the electrical operating power can also be generated on board the vehicle (e.g., by a fuel cell system and / or by a solar cell).

[0005] The term "electric vehicle" also encompasses any other vehicle in which at least some of the drive power is provided by a power battery, such as bicycles with electric drive assistance (such as electric bicycles, pedal and freight bicycles), electric scooters or electrically driven small mopeds, and in fact electrically driven wheelchairs.

[0006] Batteries with battery cells based on lithium-ion technology are currently widely used as power batteries due to their relatively high specific storage capacity (i.e., storage capacity relative to the weight of the power battery). However, there are also power batteries in use whose battery cells are based on long-established lead technology. The method described herein and its advantageous improvements are in any case not limited to a specific type of battery cell, but can instead be applied to various types of power batteries or various types of battery cells.

[0007] Like other types of rechargeable batteries, such as those used in mobile phones or other electrical or electronic devices, the power battery of an electric vehicle undergoes an aging process that is determined by both aging over time and aging due to use (i.e., the number of charge and discharge cycles). The aging process of the power battery particularly causes the storage capacity of the power battery to decline over time.

[0008] In a typical lithium-ion battery, after a service life of 3 to 4 years, approximately 80% of the original storage capacity is usually still available. After 6 years, only approximately 60% to 70% of the original storage capacity is retained. In this regard, a significant decline in capacity typically occurs over the life cycle. However, depending on the actual use of the power battery, the storage capacity can vary quite substantially. Higher temperatures, higher charge and / or discharge currents, deep discharges, or higher continuous loads result in particularly severe aging and an associated decline in storage capacity.

[0009] Figure 1 An example of the aging curve of a power battery is shown. For a power battery with a normal storage capacity Q of 70 Ah N this curve indicates the actual storage capacity Q (in ampere-hours (Ah)) as a function of the number of aging cycles. The test on which this curve is based was carried out at an ambient temperature of the power battery of 40 °C. The charge current was 1C, and the discharge current was 3C for determining the aging cycle N. In indicating 1C and 3C, the quantity C represents the charge or discharge current (in amperes (A)) with the value of the nominal storage capacity Q N of the power battery.

[0010] Although all the aging mechanisms of power batteries, especially lithium-ion storage batteries, are not yet fully understood, it is at least known that the aging process has a significant impact on the ohmic internal resistance of the power battery or battery cell and to some extent on the impedance (i.e., AC resistance).

[0011] Therefore, the actual storage capacity Q of the power battery can be measured not only by fully discharging a fully charged power battery while continuously measuring the discharge current and discharge time, but also by an alternative method in which the ohmic internal resistance of the power battery is established by applying a test load based on the output voltage of the power battery and the corresponding values established at at least one time point of the charge or discharge current. The mutual relationship between the actual capacity Q and the ohmic internal resistance R i of the power battery is in Figure 2This is shown by way of example. The test load can take the form of a load resistor (i.e., an energy absorption device or current consumer that generates a discharge current), or the form of a charging device (i.e., an energy source or current generator that generates a charging current). The load current can accordingly be a discharge current or a charging current.

[0012] Using this correlation, a battery power state value characterizing the aging state of the power battery can then be established. The actual storage capacity of the power battery can be established as the state value, for example. However, more commonly, the state value is expressed in the form of a health value, where the “health value” is used synonymously with the more common term “SoH value (SoH = “state of health”)”. The SoH value describes the ratio of the current storage capacity of the power battery (which decreases due to aging) to its original, ideal, or nominal storage capacity, and is usually described as a percentage.

[0013] Based on the output voltage and the charge or discharge current values obtained when applying a known test load, the ohmic internal resistance of the power battery mentioned above is established based on the electrical model or equivalent circuit diagram of the battery or battery cell. Figure 3 An exemplary equivalent circuit diagram showing the structure of a battery cell composed of a voltage source, various resistors R1 to R4, and various capacitors C2 to C4 is shown. Other equivalent circuit diagrams having, for example, only one series ohmic internal resistor R1 and one R-C element R2, C2 can also be used. In the prior art, for example, in the publication “Beitrag zur Bewertung des Gesundheitszustands von Traktionsbatterien in Elektrofahrzeugen”

Contribution to the Assessment of the Health State of Traction Batteries in Electric Vehicles

[0014] Surveys have shown that the described parameters depend in particular on the state of charge (SoC), state of health (SoH) and temperature of the battery. In particular, the ohmic resistance (R1) depends on the battery temperature and the state of charge (SoC). The SoC describes the current charge state of the accumulator and is usually described as a percentage value, where 100% represents a fully charged accumulator. The depth of discharge (DoD or DOD) is obtained by subtracting the state of charge value from 100%.

[0015] It has also been found that as ageing progresses due to the number of charge / discharge cycles and the consequent reduction in SoH, in particular the ohmic resistance R1 increases. For the resistances R2 to R4 and the capacitances C2 to C4, relatively small changes were observed.

[0016] When the battery temperature rises, the resistances R2 to R4 and the capacitances C2 to C4 undergo relatively small changes. In contrast, the ohmic resistance R1 decreases significantly with increasing temperature.

[0017] It has also been found that the ohmic resistance R1 remains practically constant over a wide SoC range, typically from 10% to 90%, and in particular from 30% to 90%. No significant increase in the ohmic resistance R1 was observed until the battery was discharged beyond 90%.

[0018] In summary, studies carried out on individual battery cells have shown that essentially only temperature and ageing have a significant influence on the resistance and capacitance values of the equivalent model, while the state of charge of the battery cell has little or no influence on the ohmic resistance.

[0019] It has also been found that the value of the ohmic resistance R1 is most suitable for drawing conclusions about the state of ageing of the battery. It is well known that over time, the storage capacity of a power battery decreases while the internal resistance increases. It is also well known that during a long continuous load, such as during a journey on a motorway, the internal resistance increases and with increasing continuous load, the power delivery capacity is reduced. For example, Figure 10 shows the trend of increasing internal resistance R as the power battery discharges further, where, over time, the influence of additional R-C elements (see i the equivalent circuit diagram of the battery in Figure 3 ) results in a clearly non-linear internal resistance behaviour over an extended time range, i.e. under long-term load, which can be a multiple of the internal resistance R i in the idle state. Previously used methods generally considered this total resistance when evaluating the continuous power capacity and were not suitable for determining the magnitude of the series internal resistance R1 according to the Figure 3 equivalent circuit diagram.

[0020] Conventional methods for determining the state of a power battery are characterized by an energy-intensive full discharge process. Since the error-influencing parameters discharge rate, initial discharge voltage, final discharge voltage, and temperature have to be taken into account, the determination here becomes uncertain. The conventional duration of the diagnostic determination is at least 3 hours. In order to obtain battery values such as current and voltage, vehicle accessories have to be removed in order to obtain the current connection to the physical measurement points. A continuous discharge load is also required (e.g., a rolling road test bench, a road test, or a battery test system), which is time-consuming and laborious and means a high energy input for recharging the power battery after the diagnosis is completed.

[0021] Impedance spectroscopy is an alternative diagnostic method. This involves determining the impedance of the power battery as a function of the alternating current voltage or current frequency, i.e., the AC resistance. Impedance spectroscopy is usually carried out by applying an alternating current voltage, i.e., sinusoidally modulating the potential of the working electrode and measuring its current and phase. However, this requires direct current access to the battery and a complex interpretation of the results.

[0022] For the technical investigation and / or assessment of electric vehicles, especially for the commercial assessment of used vehicles, it is also desirable to determine the state of the power battery independently of environmental influences or specific measurement conditions. Summary of the Invention

[0023] The problem solved by the present invention is to further develop a method of the type initially described in such a way that the method allows for a quick and direct establishment of the state value of the power battery, preferably the SoH value.

[0024] This problem is solved by a method having the features of claim 1. Advantageous improvements of the method are set forth in the dependent claims.

[0025] The present invention provides for establishing at least one standardized variable characterizing the power battery and, based on the established ohmic internal resistance and at least one standardized variable, establishing a standardized internal resistance of a reference value based on the standardized variable, wherein the state value of the power battery is established based on the standardized internal resistance. Thus, the ohmic internal resistance of the power battery can be standardized to one or more standardized variables. These standardized variables are in particular parameters that characterize the properties or state of the power battery in each case and can have an influence on the value of the established internal resistance. An example of a standardized variable is the temperature of the power battery, as will be explained in more detail below.

[0026] If multiple standardized variables are considered, the standardized internal resistance can also be established in multiple iterations, where in each case, in a subsequent standardization step, a corresponding standardized internal resistance is established based on the standardized internal resistance established in the previous standardization step and at least one standardized variable of the subsequent standardization step, and an additional standardized internal resistance related to the reference value of this standardized variable is established.

[0027] Therefore, the measured ohmic internal resistance is converted into an internal resistance normalized to one or more normalization variables, which enables comparability of internal resistance measurements performed on different power batteries and under different conditions by eliminating the boundary conditions for internal resistance determination caused by normalization. The normalized internal resistance can then ultimately be used to establish a state value of the power battery, such as the SoH value, where a mathematical function or a table, such as a look-up table, can be used as a basis. For this purpose, at least two value pairs of the currents I1, I2 and the voltages U1, U2 are preferably established in order to determine at least one differential current dI = I2 - I1 and at least one differential voltage dU = U2 - U1 for establishing the resistance, where the internal resistance is given by R i = dU / dI.

[0028] Overall, for establishing the internal resistance, the charging or discharging performed via a test load is designed such that the charging current or the discharging current varies according to a step change function. In order to also establish the time behavior of the change in the output voltage and / or the load current, it is also possible to provide multiple value pairs acquired temporally before the step change of the load and temporally after the step change of the load, where the time intervals can be equidistant or variable. The measured value density preferably increases directly within the load current range.

[0029] The normalized internal resistance for establishing the state value is easily established. For example, a characteristic function can be used to establish this value with a very small signal processing effort. One conceivable alternative solution (which is based on a reference performance diagram where multiple relevant parameters have to be considered) is much more complex than the solution according to the invention. In particular, the normalization method according to the invention enables a significant reduction in the effort involved in establishing the correlation between the state value and the ohmic internal resistance in the initial characterization of a specific type of battery cell. This enables a rapid and economical assessment of different types of power batteries, which can also have different aging states in terms of their state (especially their SoH).

[0030] Therefore, the state of the power battery can be established very quickly, at low cost, with low error and without disassembly within a few minutes. The method according to the invention is characterized by a low energy load on the power battery. Additionally, only compact and inexpensive test equipment is required as a diagnostic device. Due to the normalization method, only relatively minor errors are involved and independence from the energy buffer provided by the manufacturer is achieved. A direct current access for current-voltage measurement to the power battery or to the measurement point is not necessary.

[0031] The measured values can be obtained indirectly via the diagnostic interface of the vehicle, i.e., the OBD (On-Board Diagnostic) interface. The measured values can be obtained via a standard connector using a standard protocol (such as K-Line) using a serial interface or via a CAN bus. This can typically be achieved using a plug-in dongle with a wireless connection (if required) to a diagnostic device (e.g., in the form of a mobile data terminal, such as a smartphone, tablet, or laptop).

[0032] According to a preferred refinement of the method, the discharging is carried out during the evaluation run of the electric vehicle, where the test load is formed by a unit of the electric vehicle, preferably by the drive motor of the electric vehicle. The test load can be connected to initiate a charging or discharging process, e.g., by briefly accelerating strongly and then braking. The braking for establishing the internal resistance can also be omitted and is mainly carried out for test drive-specific reasons. This results in a larger load on the power battery. For example, a distance of less than 100 m, preferably up to 50 m, may be sufficient for such an evaluation run. The evaluation run can optionally be repeated one or more times. Efforts can be made here to take into account the boundary conditions that are suitable for influencing the load current (e.g., electric vehicle type, prevailing weather conditions, conditions, road state, road gradient, load in the electric vehicle, or influence of the drive motor controller) in such a way that the corresponding range of boundary conditions is specified as not being weakened or exceeded. Alternatively or additionally, the boundary conditions can also be taken into account by being included in the method in the form of one or more additional standardized variables representing the characteristics. The exposure time of the test load can be very short, in the range from 10 to 100 ms to 10 seconds. However, the test load can also last from 1 second to 30 seconds, even up to 2 minutes and particularly requires between 5 and 15 minutes. The load current can generally be the discharging current, but can also be the applied charging current.

[0033] According to a further preferred refinement of the method, the first standardized variable is the temperature of the power battery during the acquisition of the output voltage and load current values, and the reference value of the first standardized variable is the reference temperature. Thus, the internal resistance established at a specific battery temperature can be converted into a standardized internal resistance based on the reference temperature.

[0034] The second standardized variable preferably characterizes the type of power battery, and the reference value of the second standardized variable is a standardization factor that correlates different types of power batteries, where the standardization factor is specified based on at least one battery type parameter. The battery type parameter can, for example, characterize the battery type of the power battery, such as lithium-ion or lead technology, and / or can also represent the type of arrangement of the battery cells of the power battery, such as a specific number of battery cells connected in series and / or in parallel. The standardization factor is preferably obtained empirically, for example, by performing reference measurements on a given type of battery under new conditions. In principle, when establishing the standardization factor, mathematical or statistical methods, such as interpolation or extrapolation, can also be additionally used, that is, the battery type parameter or the associated second standardized factor established for a battery with a certain number of rows of parallel connections and a certain number of battery cells per row can be converted to a battery in which the number of battery cells described therein is different. The standardization factor is usually dimensionless.

[0035] In yet another standardized variable (such as the state of charge (SoC) of the power battery) can also be additionally taken into account.

[0036] According to a further preferred improvement of the method, a mathematical model or a table (preferably a look-up table or a performance graph) is used to establish a state value based on the standardized internal resistance, where the parameters or values describing the mathematical model or the table are preferably retrieved from a database. The database is preferably stored on a server, where the parameters or values are retrieved via a wireless and / or wired data link (such as a mobile data link or an Internet connection). Thus, the state value can be standardized, that is, converted into a comparable variable using a reference value.

[0037] A further advantageous development of the method provides that the first standardized variable is the temperature of the power battery during the acquisition of the output voltage and the load current values, and the reference value of the first standardized variable is a reference temperature. For this purpose, the temperature of the power battery is established, where in a first measurement step, the first ambient temperature and the first ohmic internal resistance of the power battery are established at a first time point, and in a second measurement step after a predetermined time period, the second ambient temperature and the second ohmic resistance of the power battery are established at a second time point. Based on the difference between the first and second ohmic internal resistances and the specified time period, the rate of change of the internal resistance is established. Based on the rate of change of the internal resistance, the differential temperature between the ambient temperature and the temperature of the power battery is established. The temperature of the power battery is established by adding a reference ambient temperature established based on the first and / or second ambient temperatures and the established differential temperature.

[0038] As explained above, the ohmic internal resistance of a power battery depends to a large extent on the temperature of the power battery. In many electric vehicles, direct access to the power battery for establishing the battery voltage and the charging or discharging current is not possible. These parameters are regularly provided, such as the current and voltage values at the interface of a vehicle diagnostic system (OBD, on-board diagnosis), for example in the form of pairs of values generated at a certain fixed or variable sampling rate. However, usually, the battery temperature, which can optionally be established by the vehicle diagnostic system at one or more locations, is not provided or is only provided in encrypted form. In some cases, the battery temperature can only be read out with a considerable additional time or processing effort, where usually additional manufacturer-specific requirements have to be taken into account.

[0039] The above improvement of the method provides a possible way to indirectly establish the battery temperature. This method makes use of the fact that the battery temperature deviating from the ambient temperature adapts to the ambient temperature over time, at least in the case where the battery temperature is not affected by an extended charging and / or discharging process. The adaptation of the battery temperature to the ambient temperature follows an "adaptation gradient", which is a characteristic of the battery type used and depends in particular on the design of the power battery, especially on the number and arrangement of the battery cells present in the power battery. The power battery has a rate of change of its internal resistance, i.e., the resistance difference per unit time, which depends on the difference between the battery temperature and the ambient temperature, i.e., the differential temperature. This dependence of the rate of change of the internal resistance on the differential temperature is described by a non-linear function. Based on the measurement of the rate of change of the internal resistance, the temperature difference prevailing at the time of measurement can be established. Then, the temperature of the power battery is calculated based on the reference ambient temperature and the established differential temperature, where the reference ambient temperature can be established, for example, by averaging the first and second ambient temperatures. Assuming that the ambient temperature has not changed or has not changed significantly between the two described time points, the reference ambient temperature can also be equal to the first or second ambient temperature.

[0040] The above-mentioned predetermined time period between the two measurement steps is preferably between five and fifteen minutes. During this time period, a sufficiently large change in the power battery temperature can be expected to result in a corresponding change in the ohmic internal resistance of the power battery.

[0041] In principle, the first ohmic internal resistance can be used to determine a state value of the power battery. According to a preferred improvement of the method, the state value of the power battery is established based on the second ohmic internal resistance. For this purpose, it is assumed that at the time of establishing the second ohmic internal resistance, the battery temperature will be approximately closer to the ambient temperature than in the case of the first time point. It can be seen therefrom that due to the smaller differential temperature, the rate of change of the internal resistance is also reduced, thereby improving the measurement accuracy. In addition, a recurring measurement of the ohmic internal resistance is not required, since the actual ohmic internal resistance measured at the second time point for establishing the temperature can also be used to establish the state value.

[0042] When applying a test load, the current curve of the load current changes. This can, for example, take the form of a current ramp (i.e., the charging or discharging current rises or falls linearly). According to a further advantageous refinement of the method, charging or discharging is carried out with the aid of the test load such that when the test load is connected, a step change in the load current occurs. Thus, the load current can be measured in the form of a step change response.

[0043] Preferably, a measurement sequence of output voltage and load current value pairs is acquired starting from the connection of the test load, where the measurement sequence comprises a plurality of output voltage and load current value pairs acquired at closely consecutive time points. In this way, the time curves of the output voltage and the load current can be established such that ultimately the non-ohmic component of the internal resistance, i.e., the impedance, can also be established. The impedance consists of an imaginary part (i.e., reactance) and a real part (i.e., ohmic resistance). The reactance is usually negligibly small such that the impedance can be used to estimate the ohmic resistance very precisely.

[0044] Thus, the measurement sequence represents the step change response to the generated load or current step change when the test load is connected. The time interval for acquiring the corresponding value pairs can be constant (i.e., a fixed sampling frequency or rate is provided) or can also vary, where the time density is preferably at the load step change. Optionally, the time interval can also be specified by the OBD and is not influenceable.

[0045] To ensure that the current or load step change can be optimally sampled, the measurement sequence can be acquired in such a way that the first value pair is acquired immediately at the latest when the test load is connected, where, however, the measurement sequence can also preferably be recorded a few milliseconds before the current step change. Thus, the phrase "starting from the connection of the test load" also encompasses this case.

[0046] Preferably, it is provided that before connecting the test load, at least one output voltage and load current value pair is additionally acquired, and based on this at least one output voltage and load current value pair, the open-circuit voltage U0 and the closed-circuit current I0 are established. On this basis, the ohmic internal resistance of the corresponding further value pairs of the measurement sequence can be established as the quotient of the difference between the acquired output voltage and the open-circuit voltage U0 and the difference between the acquired load current and the closed-circuit current I0. Then, the parameters of the logarithmic function for modeling the curve of the measurement sequence can be established for the measurement sequence by means of a mathematical adjustment calculation, where, based on the logarithmic function, the ohmic internal resistance can be established at least approximately at the desired time point, preferably at the current step change, or at the corresponding frequency.

[0047] The open-circuit voltage U0 or the closed-circuit current I0 is considered to refer respectively to the output voltage or the load current that predominates when the power battery is loaded with a basic load, where the basic load has a much higher ohmic resistance than the test load.

[0048] Studies have shown that immediately following the connection of the test load, the time curve of the internal resistance has the graph of a logarithmic function. By fitting the parameters of the logarithmic function to the value pairs of the measurement sequence described again, the ohmic internal resistance can be established particularly precisely, since the actually measured internal resistance can be established at the desired time point, especially with respect to the time of the current step change. Using the interpolation of the logarithmic function means that the internal resistance R i can occur at a time point very close to the load activation time t0 or at another time point t i with relatively high precision, even if no or only inaccurate data are available for this time point.

[0049] It is also particularly advantageous to use the compensation function if the time reference between the time points at which the corresponding value pairs are acquired and the time of the step change in the current applied (connection of the test load) cannot be accurately recorded. Then, for example, by introducing the time offset constant t offset , the time correlation can subsequently be established according to the graph of the logarithmic function, where by maximizing the coefficient of determination R2 with respect to the change in the offset constant t offset between the function curve and the measured values, the step change time can be estimated as precisely as possible. In this regard, the coefficient of determination R2 describes the goodness of fit of the regression, especially the goodness of fit of the regression function of the internal resistance behavior, in order to evaluate the degree to which the measured values fit the assumed model of the internal resistance.

[0050] The evaluation of the compensation function also takes into account the fact already explained at the beginning, that is, the internal resistance of the power battery has a complex value, that is, it includes a real part independent of frequency and an imaginary part related to frequency. The real part of the complex internal resistance corresponds to the ohmic internal resistance, where depending on the test conditions, the latter is usually not directly measurable. Although the imaginary part of the internal resistance is much smaller than the real part, that is, about one order of magnitude smaller, considering the imaginary part when establishing the internal resistance can improve the accuracy. Generally, the influence of the imaginary part on the total impedance can be neglected.

[0051] In principle, the internal resistance can be established not only in the form of the response to a step change, but also by periodic load changes, where the periods or frequencies of these load changes are varied. This method is called impedance spectroscopy. The results of such impedance spectroscopy are shown in Figure 13 for an exemplary power battery, where establishing the real part and the imaginary part Re(Z) and Im(Z) requires a relatively large amount of technical work. Figure 13 The so-called Nyquist plot is shown, that is, each point of the imaginary part Im(Z) and the real part Re(Z) of the complex internal resistance Z is represented at a specific frequency.

[0052] If the time curve of the internal resistance of the power battery is known for known types, ages, states of charge, temperatures, and other parameters, the position of the load activation time can be determined relatively accurately with the help of a fitted logarithmic function. Alternatively, this activation time can be estimated by varying the assumed position of this time point and maximizing the coefficient of determination R 2 To estimate this activation time.

[0053] For this purpose, the described least squares method is used to establish R from the measured curve of the battery i . The resulting internal resistance R i depends on time and can in principle be determined according to the following formula:

[0054] R i (t i ) = a·ln(t i ) + b,

[0055] where t i is the time elapsed since the load was connected, R i (t i ) is the interpolated internal resistance at time t i , and a and b are parameters. In this case, b corresponds to the internal resistance R i at a study frequency of 1 Hz or the internal resistance after applying a step change function for 1 second. The factor a approximates the effect of the RC element on the time response in the above equivalent model ( Figure 3 ).

[0056] The above dependence can also be described in a refined form as

[0057] R i (t i ) = a·ln(t i + t offset ) + b,

[0058] where is the time between the actual activation time and the suspected activation time. For this purpose, using optimization, a good estimate of t 2 can be made by maximizing the coefficient of determination R offset , where the coefficient of determination R 2 is used to evaluate the goodness of fit of the regression, i.e., the degree to which the measured values fit the underlying battery model. This allows the difference between the estimated switching time and the actual switching time to be determined and allows the actual switching time to be established. t offset can thus initially be set to zero. The coefficient of determination R 2 can be used to determine the quality of the curve. By varying t offset , the maximum R 2 can be optimized and thus the activation time estimated relatively precisely. The activation time is then at t 1Sek-t Offset -1 s. This only applies to step change excitation and has only limited use for ramp excitation.

[0059] Equations similar to the above equation can alternatively describe the internal resistance R i Dependence on the research frequency f, where the following equation applies:

[0060] R i (1 / f) = a·ln(1 / f) + b

[0061] Further optimization methods capable of achieving higher signal processing accuracy are proposed below.

[0062] According to a further preferred improvement, the expected load current is pre-determined by means of a test load, preferably by means of the ohmic resistance of the test load, and by means of the output voltage of the power battery (for example, the nominal output voltage or the open-circuit output voltage), and those value pairs in which the difference between the expected load current and the acquired load current exceeds a pre-determined tolerance value are not considered for establishing the ohmic resistance of the power battery. This allows unreliable value pairs to be excluded when establishing the state value. The tolerance value described may be, for example, 20%, preferably 10%, particularly preferably 5% of the expected load current.

[0063] A further preferred improvement provides for obtaining at least one output voltage and load current value pair of the power battery in a plurality of passes, where in each pass the test load is reconnected and removed again at the end of the pass, the corresponding value pair is acquired at at least one time point of the pass, and the corresponding ohmic resistance of the power battery is established based on the acquired value pair, and where the average value of the ohmic resistance is established based on the corresponding ohmic internal resistances established in the plurality of passes, and where the state value of the power battery is established based on the average value of the ohmic internal resistance. This repeated connection of the test load enables the accuracy of establishing the state value to be improved by averaging.

[0064] A further preferred improvement provides that establishing the ohmic internal resistance of the power battery additionally includes at least one of the following steps:

[0065] a) For at least one value of the value pair, define a corresponding effective measurement range, where if one or both values are outside the corresponding measurement range, the value pair is not considered, and where the measurement range is preferably defined based on the absolute value or rate of change of the associated value.

[0066] Technical deficiencies may mean that individual incorrect measured values are collected, which stand out due to significant deviation from the expected values. Such outliers lead to significant errors when establishing the status value. These incorrect measured values can be evaluated and screened according to fixed criteria, where various criteria can be defined for screening. If the voltage and / or current of a value pair is significantly outside the expected measurement range, the value pair can be discarded. The measured values can be evaluated both absolutely and relative to temporally adjacent measured values.

[0067] b) Obtain a measurement sequence of output voltage and load current value pairs, where the test load is connected throughout the duration of the measurement sequence, where the measurement sequence includes multiple output voltage and load current value pairs obtained at rapidly consecutive time points, and where if one or both values of a value pair are equal to the corresponding values of at least one value pair obtained at a previous time point, that value pair is disregarded.

[0068] This makes it possible to filter out, for example, "fixed" measured values, where due to errors, one or both values of a value pair do not change in one or more consecutive measurements.

[0069] c) Obtain a measurement sequence of output voltage and load current value pairs, where the test load is connected throughout the duration of the measurement sequence, where the measurement sequence includes multiple output voltage and load current value pairs obtained at rapidly consecutive time points, and where the measurement sequence is subjected to low-pass filtering.

[0070] In the case of incorrect measured values, low-pass filtering can improve the accuracy of establishing the status value, especially in cases where it is not possible to repeat the measurement multiple times.

[0071] d) Obtain a measurement sequence of output voltage and load current value pairs, where the test load is connected throughout the duration of the measurement sequence, where the measurement sequence includes multiple output voltage and load current value pairs obtained at rapidly consecutive time points, and where

[0072] d1) Establish the average value of the ohmic internal resistance based on the output voltage and load current value pairs of the measurement sequence, where, preferably, the corresponding ohmic internal resistance of two consecutively obtained value pairs is established by dividing the difference between two output voltages by the difference between two load currents, and the moving average of the ohmic internal resistance is formed by the average value of the corresponding ohmic resistances established in this way, or

[0073] d2) Establish the ohmic internal resistance based on a mathematical adjustment calculation, preferably according to the least squares method, based on the corresponding output voltage and load current value pairs.

[0074] According to variant d1, the evaluation is continuously repeated for the same measurement sequence, i.e., the ohmic resistance is established, and then the average value is formed.

[0075] As an alternative to establishing the moving average, according to variant d2, the ohmic internal resistance can be established based on a mathematical adjustment calculus, preferably according to the least squares fitting method, based on multiple value pairs. In graphical terms, for each value, the voltage is plotted against the current (and vice versa), and the best fit line is established according to the least squares fitting principle. Compared with the alternative moving average method described above, using the adjustment calculus can provide reliable results even in the case of a smaller number of value pairs.

[0076] The two methods in d) can be further improved by using additional statistical procedures, such as establishing the range, interquartile range, variance, deviation or standard deviation, and if necessary, if the data quality is too low, it can be discarded. These statistical methods can be applied to both the results and the original measurement data, i.e., the original current and voltage values.

[0077] Two or more of the above-described methods can also be combined with each other to achieve a further improvement in data quality.

[0078] If there are no directly measured value pairs of current (especially the closed-circuit current I0) and voltage (especially the open-circuit voltage U0) available, the voltage (especially the open-circuit voltage) of the power battery can be approximately determined in the voltage-current diagram. For this purpose, at least two, especially multiple measured value pairs of current and voltage, should be compared, and the best fit line is established through the measured value pairs by means of the least squares method. For example, determining the intersection of this best fit line with the current axis I (i.e., at the closed-circuit current I0) enables the estimation of the open-circuit voltage U0 of the battery in the current state of charge.

[0079] The excitation of the power battery by the load current can be in the form of a charging or discharging current. The excitation can be a ramp excitation or a step-change type excitation. If there is a step-change type excitation, the previously established open-circuit voltage U0 and closed-circuit current I0 can be used as references for determining the internal resistance R i The identification of the step-change time can be improved by optimizing the determination coefficient R of the curve fitting. Additionally, an improved identification of the time-dependent internal resistance R 2 (t i ) can be achieved. i

[0080] The invention also relates to a diagnostic device for determining the state value of a power battery of an electric vehicle, wherein the diagnostic device has an evaluation unit which is directly or indirectly coupled to the power battery and is configured to execute a method according to one of the above preferred embodiments of the invention. A direct coupling of the diagnostic device to the power battery is particularly considered to mean a coupling in which the diagnostic device can be coupled to a voltage or current measurement point of the power battery. An indirect coupling particularly includes that the diagnostic device is coupled to a diagnostic device, in particular an on-board diagnostic system (OBD) of the electric vehicle, wherein this diagnostic device of the electric vehicle transmits at least the output voltage and the load current value, and preferably also the temperature of the power battery, to the evaluation unit.

[0081] The diagnostic device can preferably access a central data repository regarding battery data, such as vehicle type, battery type, historical measurement values or comparison values of a fleet scope. The data repository can particularly be implemented as a central cloud storage on the Internet.

[0082] Advantageously, not only can the diagnostic device be regularly and independently supplied with new software updates, but also the central data repository can be regularly supplemented with measurement values and diagnostic results of various diagnostical processes distributed in time and space. The analysis based on this can also perform a type-specific or fleet-related evaluation of the SoH behavior of the power battery related to a single vehicle.

[0083] The diagnostic device advantageously includes a movable, portable, independent and wirelessly couplable display and / or input device which can be used in the scenario of test inspection inside or outside the vehicle to control the diagnostic process and display diagnostic data. In particular, the diagnostic device can take the form of a conventional mobile terminal, such as a smart phone, a tablet computer or a laptop computer. Description of the Drawings

[0084] Additional advantages are revealed by the description of the drawings of the invention. The drawings show improvements of the invention. The drawings, the description and the claims contain many features in combination. Those skilled in the art will also conveniently consider these features separately and combine them into meaningful additional combinations.

[0085] In the drawings:

[0086] Figure 1 is a graph showing the storage capacity Q of the power battery as a function of the number N of aging cycles;

[0087] Figure 2 is a graph showing the capacity Q and the internal resistance R of the power battery as a function of the number N of aging cycles i of;

[0088] Figure 3is the equivalent circuit diagram of a battery cell or a power battery;

[0089] Figure 4 is a block diagram showing the determination of state values of a power battery according to an improvement of the method according to the present invention;

[0090] Figure 5 is a graph showing the approach of the temperature T of the power battery to the ambient temperature as a function of time t;

[0091] Figure 6 is a graph indicating, by way of example for a specific battery type, the rate of change of the internal resistance of the power battery as a function of the temperature difference between the battery temperature and the ambient temperature;

[0092] Figure 7 is a block diagram indicating the determination of state values of a power battery according to a further improvement of the method according to the present invention;

[0093] Figure 8 is a graph indicating the output voltage and load current of the power battery as a function of time in the case of a step change type connection of a test load;

[0094] Figure 9 is a graph indicating the output voltage and load current of the power battery as a function of time in the case of a ramp connection of a test load;

[0095] Figure 10 and FIG. 11 are various graphs of the internal resistance of the power battery indicating the measured values as a function of time for various partial processes;

[0096] Figure 12 is a graph of the current-voltage curve I / U of a power battery having an internal resistance R i established by means of a least squares regression line;

[0097] Figure 13 is a graph indicating the real part Re(Z) and the imaginary part Im(Z) of the complex internal resistance Z of the power battery for different frequencies; and

[0098] Figure 14 is an indication of the coefficient of determination R offset for the determined internal resistance of the power battery as a function of the change in the time point t 2 for the connection of a test load; DETAILED DESCRIPTION

[0099] Figure 4It shows a block diagram showing the determination of the state value of the power battery of the electric vehicle 20 by the diagnostic device 10, which is set to execute an improved method according to the present invention. The diagnostic device 10 has an evaluation unit, which at least includes an establishment module 12 and a normalization module 14.

[0100] The diagnostic device 10 is indirectly coupled to the power battery (not shown) of the electric vehicle 20, so that the values of the output voltage U, load current I, and temperature T of the power battery can be transmitted to the evaluation unit. The coupling to the power battery can be directly carried out on the corresponding measurement points or sensors, or indirectly through the interface to the on-board diagnostic device (OBD) of the electric vehicle 20.

[0101] The power battery of the electric vehicle 20 can be coupled to a test load (such as the drive motor of the electric vehicle 20) to discharge the power battery.

[0102] The output voltage U and load current I are transmitted to the diagnostic device 10 in the form of corresponding value pairs, where the diagnostic device 10 can obtain single value pairs and measurement sequences of multiple value pairs established at specific time intervals.

[0103] The establishment module 12 is designed to determine the ohmic internal resistance of the power battery. In addition to actually calculating the ohmic internal resistance according to the output voltage U and load current I, the establishment module 12 can perform additional data processing steps, as will be explained in more detail below with special reference to further improvements of the method according to the present invention or the diagnostic device according to the present invention.

[0104] The established ohmic internal resistance is transmitted to the normalization module 14. In addition, the temperature T of the power battery is also transmitted to the normalization module 14. In the Figure 4 exemplary embodiment, the temperature T of the power battery (also abbreviated as battery temperature hereinafter) is established by a sensor system that can be arranged inside the power battery and transmitted to the normalization module 14. An alternative way of establishing the battery temperature T is explained as a variant in the exemplary embodiment to be described in more detail below.

[0105] As explained above, since the ohmic internal resistance of the power battery depends to a large extent on the battery temperature, a normalized internal resistance is established in the normalization module, which converts the ohmic internal resistance measured at the current battery temperature into a normalized internal resistance. Therefore, the battery temperature constitutes a normalization variable, where when establishing the normalized internal resistance, the currently established ohmic internal resistance is related to a reference internal resistance through a normalization function or normalization table, and the reference internal resistance is established at a reference temperature within the scope of a test procedure previously carried out on the same or also different structured power batteries.

[0106] In order to be able to take into account different battery types, a normalization factor can be considered as an additional normalization variable. The normalization factor can be provided based on at least one battery type parameter, where the battery type parameter can be, for example, the battery cell type (lithium-ion, lead, etc.), a plurality of battery cells connected in series and / or in parallel, etc. The normalization factor can be obtained empirically (e.g., by measuring the battery under new conditions) and is typically dimensionless. The state of charge (SoC) of the power battery can be considered as an additional normalization variable.

[0107] The test load can be connected, for example, by performing an evaluation run with the electric vehicle 20 over a relatively short distance, for example, up to 100 m, preferably up to 50 m, where the highest possible acceleration is advantageously set. For example, the shorter evaluation run, for example, includes short, powerful accelerations. Generally, the load usually depends on various test boundary conditions (the driving style of the tester, the vehicle, weather conditions, road conditions such as road surface or gradient, vehicle load, the function of the vehicle start control system, etc.). If necessary, the relevant test boundary conditions can be taken into account in the form of additional normalization variables.

[0108] For example, normalization can be performed based on a table, such as a look-up table or a performance graph, or it can also be based on a mathematical model. If multiple normalization variables are to be considered, normalization (i.e., establishing the normalized internal resistance) can also be performed in multiple sub-steps.

[0109] The parameters required for normalization (i.e., the parameters of the performance chart or performance graph values or the mathematical model or the mathematical normalization function) can be stored in the diagnostic device 10 and / or can also be retrieved by the diagnostic device 10 from an external database. The diagnostic device 10 can also optionally feedback the corrected values of these parameters back to the database.

[0110] The normalization module 14 is also designed to establish a state value of the power battery based on the normalized internal resistance. In the present exemplary embodiment, the state of health (SoH) of the power battery is output as the state value. For example, the SoH can be calculated based on an SoH assignment function or table established empirically in previous tests. The SoH is output, for example, in the form of a log printout 16. It goes without saying that there are also other output options, such as by means of a display or also by means of wireless or wired transmission to a suitable display, acquisition, or data processing device. For example, the SoH value together with other acquired parameters (such as the input parameters U, I, and T and optional vehicle identification data) can be transmitted to a central server, from which it is transmitted in paper or electronic form, for example, to the user of the electric vehicle or to a workshop.

[0111] One advantage of such a central data storage device is that different power batteries or electric vehicles (even those that were not initially tested) can be analyzed. For example, a standard consisting of empirical values from multiple batteries or electric vehicles can be defined for several similar power batteries or electric vehicles. Compared to referencing a single reference battery, the influence of any manufacturing tolerances can thus be reduced, and given a suitably large database, the corresponding aging state of the power battery can also be taken into account.

[0112] Reference Figures 5 to 7 A further advantageous improvement of the method or diagnostic device is described. The measured values of the output voltage U and the load current I required to establish the internal resistance of the battery are usually provided by the OBD of the electric vehicle or another interface, because access to components at high voltage inside the battery is usually not possible for safety reasons. However, depending on the specific manufacturer, the battery temperature T is often not provided at all or only in encrypted form. To ensure the use of the diagnostic device or method as independent of the manufacturer as possible, this advantageous improvement provides an alternative method for establishing the battery temperature.

[0113] This makes use of the fact that when determining the state value, the battery temperature is different from the temperature of the environment in which the power battery test is carried out. In the exemplary case forming the basis for Figure 5 it is assumed that the electric vehicle is parked outside, such that the power battery has a temperature T of only approximately 4 °C at the start of the test. On the other hand, the test is carried out in a building where the ambient temperature of the room temperature or approximately 24 °C prevails. Thus, over time t, the battery temperature will rise from 4 °C to the ambient temperature of 24 °C. This process is non-linear and is represented by the solid line in Figure 5 .

[0114] Two parts of this adaptation curve are additionally marked by two corresponding pairs of circles in Figure 5 . It is evident from Figure 5 that the rate of temperature change ΔT / t represented by the gradient of the curve is relatively high at the start of the adaptation process (the part of the curve enclosed by the left pair of circles ΔT / t1), that is, where there is a large temperature difference; and decreases with increasing adaptation, that is, with the decreasing temperature difference between the environment and the battery (the part of the curve enclosed by the right pair of circles ΔT / t2). Based on the known fact that the ohmic internal resistance of the power battery varies as a function of the battery temperature, the rate of change of the battery temperature can be established based on the change in internal resistance per unit time (i.e., the rate of change of the internal resistance ΔT / t).

[0115] If Figure 5The temperature scale on the y-axis is now standardized such that the value that the curve approaches over time is considered to be the differential temperature with a value of 0 °C, and the curve then directly indicates the differential temperature between the ambient temperature and the battery temperature. Based on the rate of change of the internal resistance, the rate of change of temperature can be established, and from the rate of change of temperature (which, as explained above, corresponds to the gradient T / t of the curve in Figure 5 ), the associated differential temperature can in turn be derived by appropriate computational steps. Since the ambient temperature is known or can be easily measured, the battery temperature can be directly inferred based on the ambient temperature and the differential temperature.

[0116] The rate of change of the internal resistance over time, ΔR i / t, and the correlation with the differential temperature ΔT between the surroundings and the battery can also be represented directly in a simplified manner by a corresponding curve ΔR i / t plotted against ΔT, see Figure 6 . Based on this curve, which can be represented by a mathematical model or a table, the differential temperature can be directly read off for a given rate of change of the internal resistance.

[0117] The rate of change of the internal resistance is determined by determining the first ohmic internal resistance and the second ohmic internal resistance of the traction battery, for example, at intervals of 5 to 15 min. The difference between the first and second ohmic internal resistances, divided by the time interval between the two measurements, then gives the rate of change of the internal resistance, ΔR i / t. At the same time as measuring the ohmic internal resistance, the ambient temperature T can also be recorded, and the battery temperature can then be established from this ambient temperature by adding the differential temperature. If the ambient temperature T varies slightly during the measurement, the average value of one of the two ambient temperatures or the ambient temperature T established at different times can be used as the reference ambient temperature. Advantageously, more than two internal resistances R i can be established to obtain a higher accuracy for determining the rate of change of the resistance.

[0118] Figure 7 Fig. shows a diagnostic device 110 that is designed to carry out this variant of the method according to the invention. Since the diagnostic device 110 is a Figure 4 variant of the diagnostic device 10 shown in

[0119] only the significant differences will be described below. Identical or similar elements therefore have the same reference numerals. i The diagnostic device 110 additionally has a buffer module 18 in which the internal resistance values R U established by the establishing module 12 can be temporarily stored.

[0120] At a first time point, an internal resistance R is established based on the output voltage U and the load current I i , and it is transmitted to the buffer module 18 and temporarily stored there as the internal resistance value R i1 . Once a predetermined time period, for example, between 5 and 15 minutes, has elapsed, the output voltage U and the load current I are measured at a second time point and converted into a further internal resistance value R in the establishment module 12 i2 . This second internal resistance value is transmitted to the normalization module 14. The first internal resistance value R temporarily stored i1 is simultaneously transmitted from the buffer module 18 to the normalization module 14. The normalization module now establishes the difference between the two internal resistance values R i1 、R i2 and divides this difference by the time period between the two measurement times.

[0121] Based on the rate of change of the internal resistance established in this way, Figure 6 the battery-specific curve (which can be stored, for example, in the form of a table or a mathematical function in the diagnostic device 110) is used as the basis for establishing the associated differential temperature. Based on this differential temperature and the ambient temperature T u , the normalization module 14 now establishes the battery temperature, which can then be used as a reference Figure 4 for the normalization described.

[0122] Now a further improvement and variation of the method according to the invention or of the diagnostic device 10, 110 according to the invention will be described with reference to Figures 8 to 12 .

[0123] Figure 8 Shows a graph of the output voltage U and the load current I of an exemplary power battery (which reflects a step change in the charging current). In each case Figure 8 the illustration shows the measured values of U and I at the corresponding time point t. The origin of the time scale is arbitrarily selected. At the step change time t0 of approximately t = 6.5 s, a test load in the form of a charging current is connected, such that the output voltage U and the charging current as the load current I change suddenly with the step change S. This time period extends approximately between t = 6.5 s and t = 7.5 s. Thereafter, U and I change only very slowly. The descriptive function of U and I can be identified by using a regression function (for example, an exponential function). If a time offset t offset is provided in this regression function, then the time base offset can be maximized by modifying t offset and the activation time t0 of the step change S established thereby to maximize the determination coefficient R 2 , for example, as shown for the respective values of t Figure 14 in offset .

[0124] Figure 9 Shows a graph of the output voltage U and load current of a power battery as a function of time in the case of a ramp-connected R with a test load (in this case, the test load is a discharge current). During the time period from t1 = 3 s to t2 = 6 s, the discharge current I gradually decreases, and after the load is terminated, the current I and voltage U swing back towards the closed-circuit current I0 and open-circuit voltage U0. Preferably, the determination of the internal resistance is improved by preferably considering the measured values within the time interval t1 to t2 of the load period of the ramp connection R, which can be approximated by a regression line, where the internal resistance R i can be accurately determined according to the regression line. To be able to accurately determine this time period, a regression function of the load current I corresponding to the test load function can be assumed, and the time period t1 to t2 can be accurately determined by considering the maximization graph of the determination coefficient R 2 .

[0125] The corresponding internal resistance value can be established based on the measured values established at a given time point. In Figure 10 the figure, these internal resistance values R i are plotted against time t, where the internal resistance values R i 1 to R i 3 are marked accordingly. Figure 10 The dashed line shown in is the best-fit line of the measured internal resistance values R i .

[0126] From Figure 10 and from Figure 11a it is clear that, in particular, the measured values R i 1 and R i 2 deviate significantly from the best-fit line, thus emphasizing the need for frequent authenticity checks. The minimized value of the determination coefficient R 2 also indicates this. To determine the internal resistance of a power battery, generally, the system response of the power battery to a system excitation is evaluated. A current is applied as the system excitation, and the system response is the voltage change at the battery terminals.

[0127] Under real conditions, the current and voltage difference during the acceleration process are used to calculate the internal resistance R i . Using the values U0, I0 in the system stationary state, i.e., the open-circuit voltage U0 and closed-circuit current I0, and the values U i , I i , I i at the time point t i , the internal resistance R

[0128]

[0129] Rather than using the drive motor to accelerate the vehicle, another energy-intensive consumer can be activated. In principle, battery charging can be provided as a system excitation.

[0130] It is generally not possible to synchronize the time scales of the system excitation (i.e., the connection of the test load or charging current) and the system response (i.e., the acquisition of the measured values). However, in order to reliably determine the real part of the internal resistance, i.e., Figure 3 the resistance R1 in the equivalent circuit diagram of, it is necessary to determine the internal resistance directly after activating the test load or at a defined point in time in the near future. Due to the large rate of change of the current and voltage, the corresponding measured value pairs are usually affected by large errors at this time or are not established at the correct time due to discrete sampling. In order to ensure the correct establishment of R1, the measurement data is therefore preferably interpolated or extrapolated. Therefore, proper interpretation of the measurement data is crucial.

[0131] A method for evaluating measurement data is explained with reference to FIG. 11, where in Figures 11a to 11d each of, the internal resistance R i is plotted against time t. The points represent the corresponding measured values, while the corresponding compensation function and the associated determination coefficient R 2 are indicated in the figure.

[0132] Figure 11a Basically corresponds to Figure 10 and shows a relatively low determination coefficient R 2 , i.e., a relatively inaccurate determination of the internal resistance curve.

[0133] In contrast to Figure 11a , in Figure 11b , since the measured value R i 1 deviates significantly from the best-fit line or the specified expected value, it is discarded. The comparison with Figure 11a shows that Figure 11b the gradient of the compensation function of 2 increases slightly, and the value of the determination coefficient R Figure 11a increases from 0.9249 ( Figure 11b ) to 0.9665 (

[0134] Figure 11c The graph shown in Figure 11b uses the same measured values as in Figure 11a and Figure 11b , however, where a logarithmic compensation function is used instead of the linear compensation function for 2 and

[0135] Figure 11d corresponding to Figure 11c , where the time axis (x-axis) has been logarithmically scaled.

[0136] Figure 12 shows a graph of the linear current and voltage curve I / U of a power battery with a ramp load R. The internal resistance R i is established by using an adjustment calculation tool to determine the regression line by the least squares method (i.e., the least squares fitting method). The determination coefficient R that can be established 2 indicates whether the compensation function (in this case a curve) under consideration fits the measured values established. This method is even very suitable for establishing the open-circuit voltage U0 with fluctuating current.

[0137] This shows that by adjusting the measurement sequence through mathematical adjustment calculations using a logarithmic function, the graph of the measurement sequence can be very accurately simulated. The logarithmic function established in this way can then be used to extrapolate or interpolate the ohmic internal resistance at a desired time point, particularly at a time point close to the start of the step change response.

[0138] When establishing the ohmic internal resistance by disregarding those value pairs for which the difference between the expected load current and the acquired load current exceeds a predetermined tolerance limit (as done for the value pairs on which the internal resistance R i 1 is based in this example), a further significant improvement in data evaluation can be achieved.

[0139] Other methods for improving the accuracy of internal resistance determination and the accuracy of state value determination are briefly described below:

[0140] 1. The internal resistance can be determined in multiple passes, where in each pass, a test load is connected and removed again at the end of the pass. At least one value pair, preferably a measurement sequence, is acquired in each pass. Then the corresponding ohmic internal resistance of the power battery is established for each pass. Finally, the average value of the ohmic internal resistance is established based on the ohmic internal resistances established in multiple passes. This average value is then used to establish the state value of the power battery.

[0141] 2. For at least one value of the value pair, a corresponding valid measurement range is defined, where if one or both values are outside the corresponding measurement range, the value pair is disregarded, and the measurement range is preferably defined based on the absolute value or rate of change of the associated value.

[0142] This can be achieved by defining corresponding valid measurement ranges for the measured values used to identify errors. The measurement ranges can be defined absolutely, for example by absolute limit values of voltage; or relatively, for example by boundaries of the voltage change rate. Subsequently, all pairs of measured values are verified as to whether the values of the pairs are within the previously defined measurement ranges, and if they do not fall within the defined measurement ranges, they are filtered out. For example, this can be done by establishing the R 2 variation due to ignoring the measured values to analyze the coefficient of determination R 2 of the values within the step change range.

[0143] 3. Obtain a measurement sequence of pairs of output voltage and load current values, where a test load is connected throughout the duration of the measurement sequence, and the measurement sequence includes multiple pairs of output voltage and load current values obtained at closely consecutive time points. If one or both values of a pair are equal to the corresponding values of at least one pair obtained at a previous time point, that pair is not considered.

[0144] This method makes it possible to eliminate "fixed" measured values that arise particularly due to transmission errors and especially delays during data transmission from the power battery or OBD. These data can lead to errors in the evaluation results, and it is very difficult or impossible to identify these errors retrospectively. However, such "fixed measured values" can be filtered out on a strict criterion basis because the main characteristic of these error data is that the measured values do not change over a certain period of time. Due to the characteristic feature of the method according to the invention (according to which the system response to a sudden system excitation will be identified), any measured value that is not different from the previously measured values can be filtered out. For example, if at least one value (i.e., voltage or current) does not change from one pair of measured values to the next within the measurement sequence, the corresponding pair of values obtained at a later time point can be deleted from the dataset to be evaluated. However, if the error of the fixed measured value only involves one value of the pair, that value can also be replaced by interpolation based on other corresponding values.

[0145] 4. Obtain a measurement sequence of pairs of output voltage and load current values, where the test load is connected throughout the duration of the measurement sequence. The measurement sequence includes multiple pairs of output voltage and load current values obtained at closely consecutive time points. The measurement sequence is subjected to low-pass filtering. With this simple filtering, a surprisingly good improvement in the accuracy of the determination of the internal resistance can generally be achieved.

[0146] Low-pass filtering is particularly useful when it is not possible to repeat multiple measurement runs (see Method 1).

[0147] A low-pass filter can be used as if multiple repeated measurements were made at similar time points, but taking into account insufficient measurement resolution and thus forming a virtual average value. This method can preferably be used for test arrangements that do not include step changes but are continuous.

[0148] 5. Here too, a measurement sequence of output voltage and load current value pairs is acquired, where the test load is connected throughout the duration of the measurement sequence. Here too, the measurement sequence includes multiple output voltage and load current value pairs acquired at time points in rapid succession.

[0149] a) Establish a moving average of the ohmic internal resistance based on the output voltage and load current value pairs of the measurement sequence. This method is particularly suitable for ramp excitation, i.e., continuous load increase or decrease. This is preferably done because the corresponding ohmic internal resistance of two consecutively acquired corresponding value pairs is established by dividing the difference between two output voltages by the difference between two load currents, and the moving average of the ohmic internal resistance is formed by the average of the corresponding ohmic resistances established in this way. Thus, in each case, the associated internal resistance is established based on two adjacent value pairs of the measurement sequence, and then the average of these internal resistances is calculated according to the following equation:

[0150]

[0151] where n is the number of value pairs, and the corresponding ohmic resistance R is established based on two consecutively acquired value pairs according to the following equation im

[0152]

[0153] where U m 、U m+1 are the corresponding output voltages, and I m 、I m+1 are the corresponding load currents of two consecutively acquired value pairs m, m + 1 in the measurement sequence. If the open-circuit voltage U0 and short-circuit current I0 are selected instead of U m and I m , this method can also be used for step-change type excitation.

[0154] The number n does not necessarily represent the number of all measured value pairs of the measurement sequence, but can also be the number of measured value pairs to be considered, for example, in the case where one or more measured value pairs have been removed from the measurement sequence or are not considered.

[0155] b) As an alternative to method 5a), the ohmic internal resistance can be established based on mathematical adjustment calculus, preferably according to the least squares method, based on the corresponding output voltage and load current value pairs. This method has been described above with reference to Figures 8 to 14 as follows.

[0156] In the absence of the above methods for improving the accuracy of determining the internal resistance and the state value, a significantly greater scatter of the result data can be observed. A variation of up to about 20% of the internal resistance was established in a series of different test runs. Applying the described method makes it possible to reduce it to 3% or less.

[0157] If the time response of the battery as a function of the battery type, age, state of charge, temperature and optionally other parameters is known, the switching time t0 at which the current function starts can be determined relatively precisely using the comparison curve established for the reference battery.

[0158] If this is not possible, the time t0 can vary when establishing the compensation function. Then the switching time t0 can be estimated by maximizing the determination coefficient R 2 of the compensation function. Figure 14 shows a plot in which the determination coefficient R 2 is plotted against the change or variation of the switching time t0 (implicitly the time offset t offset ).

[0159] List of reference numerals

[0160] 10, 110 Diagnostic device

[0161] 12 Establishment module

[0162] 14 Normalization module

[0163] 16 Log print output

[0164] 18 Buffer module

[0165] 20 Electric vehicle

Claims

1. A method for determining a state value of a power battery of an electric vehicle (20), the state value characterizing an aging state of the power battery, wherein, The power battery is loaded through a test load, and at at least one time point, a corresponding output voltage and load current value pair of the power battery is obtained, wherein an ohmic internal resistance of the power battery is established based on the obtained output voltage and load current value pair, and wherein a state value of the power battery is established based on the established ohmic internal resistance, at least one standardized variable characterizing the power battery is established and a standardized internal resistance based on a reference value of the standardized variable is established based on the established ohmic internal resistance and the at least one standardized variable, wherein the state value of the power battery is established based on the standardized internal resistance, characterized in that the test load is carried out in the following manner: when connecting the test load, the load current has a step change or a ramp curve of current, wherein a measurement sequence of the output voltage and the load current value pair is obtained starting from connecting the test load, wherein the measurement sequence includes a plurality of output voltage and load current value pairs obtained at rapidly consecutive time points, wherein parameters of a compensation function for modeling the curve of the measurement sequence are established by a mathematical adjustment calculation for the determination of the internal resistance, wherein the optimization calculation of the compensation function is carried out by maximizing a coefficient of determination R 2 that describes the goodness of fit of the adjustment calculation.

2. The method according to claim 1, characterized in that, The load current is generated during the evaluation run of the electric vehicle (20), wherein the test load is formed by a unit of the electric vehicle (20).

3. The method according to claim 1, characterized in that, The first standardized variable is the temperature of the power battery during the acquisition of the output voltage and load current values, and the reference value of the first standardized variable is the reference temperature.

4. The method according to claim 3, characterized in that, The second standardized variable characterizes the type of the power battery, and the reference value of the second standardized variable is a standardized factor that correlates different types of power batteries, wherein the standardized factor is specified based on at least one battery type parameter.

5. The method according to claim 1, characterized in that, Based on the standardized internal resistance, a mathematical model or a table is used to establish the state value.

6. The method according to claim 5, characterized in that, The table is a look-up table or a performance graph, wherein the parameters or values describing the mathematical model or the table are retrieved from a database.

7. The method according to claim 1, characterized in that, The first standardized variable is the temperature of the power battery during the acquisition of the output voltage and load current values, and the reference value of the first standardized variable is the reference temperature, wherein establishing the temperature of the power battery lies in that, in a first measurement step, the first ambient temperature and the first ohmic internal resistance of the power battery are established at a first time point, in a second measurement step after a predetermined time period, the second ambient temperature and the second ohmic internal resistance of the power battery are established at a second time point, and based on the difference between the first ohmic internal resistance and the second ohmic internal resistance and the specified time period, the rate of change of the internal resistance is established, based on the rate of change of the internal resistance, the differential temperature between the ambient temperature and the temperature of the power battery is established, and the temperature of the power battery is established by adding a reference ambient temperature established according to the first ambient temperature and / or the second ambient temperature and the established differential temperature.

8. The method according to claim 7, characterized in that, The range of the predetermined time period is from 5 minutes to 15 minutes.

9. The method according to claim 7 or 8, characterized in that, The state value of the power battery is established based on the second ohmic internal resistance.

10. The method according to claim 1, characterized in that, Before connecting the test load, at least one pair of output voltage and load current reference values is additionally acquired, the open-circuit voltage and the closed-circuit current are established based on the at least one pair of output voltage and load current reference values, the ohmic internal resistance of the corresponding value pair of the measurement sequence is established as the quotient between the difference between the acquired output voltage and the open-circuit voltage and the difference between the acquired load current and the closed-circuit current, a parameter of a logarithmic function that models the graph of the measurement sequence is established for the measurement sequence through mathematical adjustment calculus, and based on the logarithmic function, the ohmic internal resistance is established at a desired time point.

11. For the method according to claim 10, the desired time point is at a step change of the current or is allocated at a corresponding frequency.

12. The method according to claim 10 or 11, characterized in that, The logarithmic function is determined by the following equation ， where t i is the time period elapsed since the connection of the test load, R i (t i ) is the interpolated internal resistance at time t i , is the time period between the actual activation time and the estimated activation time, and a and b are parameters.

13. The method according to claim 1, characterized in that, Through the test load and through the output voltage of the power battery, the expected load current is pre-determined, wherein those value pairs for which the difference between the expected load current and the acquired load current exceeds a pre-determined tolerance value are not considered for establishing the ohmic internal resistance of the power battery.

14. The method according to claim 1, characterized in that, Obtain at least one pair of output voltage and load current values of the power battery in multiple passes, wherein, in each pass, reconnect and remove the test load at the end of the pass, obtain the corresponding value pairs at at least one time point in the pass, and establish the corresponding ohmic internal resistance of the power battery based on the obtained value pairs, and wherein, establish the average value of the ohmic internal resistance according to the corresponding ohmic internal resistances established in the multiple passes, and wherein, establish the state value of the power battery based on the average value of the ohmic internal resistance.

15. The method according to claim 1, characterized in that, Establishing the ohmic internal resistance of the power battery further includes at least one of the following steps: - Define a corresponding effective measurement range for at least one value in the value pair, wherein, if one or both values are outside the corresponding effective measurement range, the value pair is not considered. - Obtain a measurement sequence of output voltage and load current value pairs, wherein the test load is connected throughout the duration of the measurement sequence, and wherein the measurement sequence includes a plurality of output voltage and load current value pairs obtained at closely consecutive time points, and wherein, if one or both values of the value pair are equal to the corresponding values of at least one value pair obtained at a previous time point, the value pair is not considered. - Obtain a measurement sequence of output voltage and load current value pairs, wherein the test load is connected throughout the duration of the measurement sequence, and wherein the measurement sequence includes a plurality of output voltage and load current value pairs obtained at closely consecutive time points, and wherein the measurement sequence is subjected to low-pass filtering. - Obtain a measurement sequence of output voltage and load current value pairs, wherein the test load is connected throughout the duration of the measurement sequence, and wherein the measurement sequence includes a plurality of output voltage and load current value pairs obtained at closely consecutive time points, and wherein - Establish a moving average value of the ohmic internal resistance according to the output voltage and load current value pairs of the measurement sequence, wherein establish the corresponding ohmic internal resistance of two correspondingly obtained value pairs that are obtained immediately consecutively by dividing the difference between two output voltages by the difference between two load currents, and the moving average value of the ohmic internal resistance is formed by the average value of the corresponding ohmic internal resistances established in this way, or - Establish the ohmic internal resistance based on a mathematical adjustment calculation according to the corresponding output voltage and load current value pairs.

16. The method according to claim 15, wherein the measurement range is defined based on the absolute value or the rate of change of the associated value.

17. A diagnostic device (10, 110) for determining a state value of a power battery of an electric vehicle (20), wherein, The diagnostic device (10, 110) has an evaluation unit, which is directly or indirectly coupled to the power battery and is configured to perform the method according to claim 1.

Images