EP3847467A1

EP3847467A1 - Method and device for the diagnosis of battery cells

Info

Publication number: EP3847467A1
Application number: EP19783454.2A
Authority: EP
Inventors: Hendrik Zappen; Alexander GITIS
Original assignee: Rheinisch Westlische Technische Hochschuke RWTH
Current assignee: Rheinisch Westlische Technische Hochschuke RWTH
Priority date: 2018-09-26
Filing date: 2019-09-26
Publication date: 2021-07-14
Also published as: US20220011373A1; WO2020064932A1; US11567140B2; CN113196072A

Abstract

The invention relates to a method for determining a battery state of at least one battery cell comprising the steps: a. applying a current-exciting signal; b. recording an impedance spectrum of the battery cell; c. determining an evaluation variable on the basis of a measured impedance spectrum. This method is distinguished in that at least one of the parameters amplitude, frequency and relative phase difference of at least one component of the current-exciting signal is modified as a function of the first measured response signal such that a measuring error is minimized, in that a further measured response signal is determined and evaluated, in that the value evaluated is used as an evaluation variable and in that the battery state variables of the battery cell are determined on the basis of a comparison of at least one diagnostic variable with at least one reference value and/or with at least one further diagnostic variable. The invention further relates to an exciter unit for performing the method, a battery system comprising an exciter unit of this nature, a device for performing the method and a battery system comprising a device of this kind.

Description

Method and device for diagnosing battery cells Field of the invention

The invention relates to a method for diagnosing critical changes, in particular pressure changes, gas developments and temperature changes in

Battery cells, on an excitation unit for performing the method, on a battery system with such an excitation unit, on a device for performing the method and a battery system with such a device.

Background of the Invention

The use of electrical or electronic devices, especially portable devices, is often dependent on electrochemical battery cells as a power supply, so-called "battery cells" or "batteries". Batteries can be used in electronic devices such as telecommunication devices (e.g. cell phones, tablets, computers), means of transport (e.g. cars, airplanes, boats), but also in non-portable devices such as storage systems for central and decentralized power supplies.

This shifts due to the external operating conditions of battery cells, such as excessive current levels during charging and discharging, excessive mechanical forces on the cell housing or on the contacts, overcharging or deep discharging or excessive and low cell voltage, or excessive and low cell temperatures electrochemical equilibrium in the battery cell and the battery cell can get into a critical battery state. In addition, the battery cell can also get into critical states due to aging-related degeneration mechanisms. The critical battery condition can lead to loss of capacity, an increase in internal resistance, or to exothermic processes. These in turn can pose dangers not only for the devices, but also for people. Early detection is therefore desirable to

Take countermeasures before these batteries pose a danger.

The use of electrochemical impedance spectroscopy as a diagnostic method for recording operating parameters (state of charge (SOC), capacitance and internal resistance (SOH), temperature and others) or safety parameters (e.g. detection of lithium plating in lithium-based battery cells or gas formation) presupposes that the method in almost any

Operating states of a multi-cell battery storage system can be used, and also enables continuous monitoring. The prior art of impedance spectroscopy is exemplified in DE 10 2009 000 337. This method leads to unusable results if this is to be carried out during the dynamic operation of the battery with different charging and discharging currents.

Summary of the invention It would therefore be desirable to have a diagnostic facility available that does not damage the battery cell and does not have the disadvantages of the prior art.

According to the invention, this object is achieved in that a method for determining a battery state of at least one battery cell comprises the steps a) applying a current excitation signal to the battery cell; b) recording a response measurement signal and determining the impedance spectrum of the

Battery cell or the impedance of the battery cell; c) determining at least one evaluation variable based on the measured impedance spectrum or the impedance; a. such that the current excitation signal consists of at least two periodic signals with mutually different frequencies, that the current excitation signal is applied in such a way that it is free of mean values over at least one period of the smallest frequency contained, that a first response measurement signal from the battery cell b. or at least one frequency and determining at least one

Diagnostic quantity from the impedance or whereby it is determined and evaluated in such a way that at least one of the parameters amplitude, frequency and relative phase position of at least one component of the current excitation signal is changed as a function of the first response measurement signal, so that a measurement error is minimized, that a further response measurement signal is determined and it is evaluated that the evaluated value is used as an evaluation variable and that and that the battery state variables of the battery cell are determined using a

Comparison of at least one diagnostic variable with at least one reference value and / or with at least one further diagnostic variable is carried out.

The term battery status means in particular an operational, safety, or

State of aging, in particular State of Charge (SOC), State of Health (SOH), State of Function (SOF) or State of Safety (SOS) and / or temperature control of a battery.

In addition to a single cell, the term battery cell also includes macro battery cells which consist of a plurality of individual battery cells connected in parallel or in series but encapsulated together and / or having a common power connection.

The response measurement signal is a voltage response measurement signal.

The feature that at least one of the parameters amplitude, frequency and relative phase position of at least one component of the current excitation signal as a function of the first Response measurement signal is changed means in particular that the change, or the changes depending on at least one of the evaluation variables of the first

Response measurement signal take place.

A spectrum is in particular also a non-continuous spectrum, but rather a spectrum based on a finite number of values at a finite number of

Frequencies understood, the finite number of frequencies being in particular at least 2 and / or less than 100, preferably at least 5 and / or a maximum of 50, the

Frequencies are in particular linearly or logarithmically distributed, or in a quasi-logarithmic or quasi-linear distribution and / or in particular distributed such that the frequencies are chosen such that in particular the 2nd and 3rd harmonics of each of the frequencies do not coincide with other ones selected frequencies overlap. In particular, the frequencies are in a range from 1 Hz to 10 kHz.

The at least one evaluation variable is in particular one or more of the following variables:

Result of a consistency check of recorded impedance spectra using the Kramers-Kronig relationship, in the form of a residual between the measured and calculated impedance spectrum, a Kramers-Kronig residual

Signal-to-noise ratio between response measurement signal and superimposed noise signals, amplitude of the response measurement signal

Harmonic content of the response measurement signal (this is used to assess the linearity of the measurement).

On the basis of these evaluation variables, an adaptation of the current excitation signal is carried out in particular, in particular only when at least one or one of the selection of the following adaptation criteria is fulfilled, the signal-to-noise ratio (SNR) of the response measurement signal for one and / or several and / or all the frequencies contained are less than a predetermined value in the range from 1 to 25 dB, in particular less than a predetermined value in the range 15 to 25 dB. the amplitude of the response measurement signal is greater than a predetermined value in the range from 5 to 60 mV, in particular greater than a predetermined value in the range from 5 to 20 mV. the harmonic content is greater than a predetermined value in the range from 0.5 to 8%, in particular greater than a predetermined value in the range from 0.5 to 3%. the Kramers-Kronig residual is greater than a predetermined value in the range from 0.2 to 3%, in particular greater than a predetermined value in the range from 0.2 to 0.8%, which suggests a violation of the time invariance condition.

Furthermore, it is expedient to iteratively change only two of the three parameters amplitude, frequency and relative phase position.

Even if no adaptation is required according to the above adaptation criteria, a further optimization of the excitation current signal can be carried out in order to e.g. Further

Add frequency components and / or in particular by adding

Signal components with other frequencies to get a more complete impedance spectrum.

Frequency point means in particular a frequency.

The adaptive control for adapting the current excitation signal adjusts the current excitation signal in particular as follows on the basis of the evaluation variables determined:

If the SNR of a frequency point is too low, in particular less than that specified in the predetermined adaptation criterion (s), increase the excitation amplitude of the signal of this frequency point, in particular by a value in the range from 5 to 30%, if the amplitude of the response measurement signal is not yet too large , and the linearity criteria are met. Otherwise: Either delete the signal component of this frequency point from the excitation signal or optimize the phase positions of the signal components using suitable non-linear optimization methods (for example, like downhill simplex

Optimizer) in order to obtain a lower crest factor of the current excitation signal.

If the SNR of a frequency point is too high: Reduce the excitation amplitude of the signal component of this frequency point, in particular by a value in the range from 5 to 30%

If the amplitude of the response measurement signal is too high: reduce the amplitudes of the signal components, in particular evenly, or starting with the signal component with the best signal-to-noise ratio, in particular by a value in the range from 5 to 30%, or remove individual signal components

If the harmonic content is too high: Reduce the excitation amplitude of all signals, in particular by a value in the range of 5 to 30% and / or optimize the phase positions of the signal component in order to obtain a lower crest factor of the current excitation signal. If the Kramers-Kronig residual is too high: Remove the signal component of the lowest frequency from the current excitation signal.

If none of the matching criteria is met: advantageously add further signal components with different frequencies in order to obtain a more complete impedance spectrum, in particular until the number of signal components is in the range from 10 to 30.

The diagnostic variable is, in particular, one or more of or determined using one or more of the following variables: characteristic variables of the impedance such as real part, imaginary part, amount and phase of an individual battery cell and / or the gradient of one or more of these variables over at least two, in particular a plurality of (excited) frequencies, at least two impedances then having to be determined, in particular a plurality; statistical values of the impedance, in particular mean value and scattering measures for the parameters real part, imaginary part, amount and phase of a plurality of battery cells connected in series and / or in parallel;

Comparison of characteristic quantities of the impedance such as real part, imaginary part, amount and phase of an individual battery cell with quantities from the same time step for an individual battery cell or for several battery cells connected in series and / or in parallel;

Comparison of characteristic quantities of the impedance such as real part, imaginary part, amount and phase of an individual battery cell with quantities from at least one previous one

Time step for a single battery cell or for several battery cells connected in series and / or in parallel;

Comparison of characteristic quantities of the impedance such as real part, imaginary part, amount and phase of an individual battery cell with quantities from at least one beforehand

measurement carried out for a single battery cell or for a plurality of battery cells connected in series and / or in parallel;

Comparison of statistical quantities of the impedance, in particular mean value and scattering measures for the parameters real part, imaginary part, amount and phase with quantities from the same time step for a single battery cell or for several battery cells connected in series and / or in parallel;

Comparison of statistical quantities of the impedance, in particular the mean value and the scatter measures for the parameters real part, imaginary part, amount and phase with quantities from at least a previous time step for a single battery cell or for several battery cells connected in series and / or in parallel;

Comparison of statistical quantities of the impedance, in particular mean value and scatter measurements for the parameters real part, imaginary part, amount and phase with quantities from at least one previously carried out measurement for a single battery cell or for several battery cells connected in series and / or in parallel; characteristic quantities of the relaxation time spectrum, such as the value of the global one

Maximums, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference in relaxation times with respect to the maxima of a distribution function of polarization resistances , the mean of the maxima, the center of gravity of the spectrum, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several defined frequency ranges and the mean value of the integrals over one or more defined frequency ranges; statistical variables of the relaxation time spectrum, in particular mean and

Scattering measures for the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference in relaxation times the maxima of a distribution function of

Polarization resistances, the mean of the maxima, the center of gravity of the spectrum, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a certain frequency range, the integral values over several delimited

Frequency ranges and the mean value of the integrals over one or more defined frequency ranges;

Comparison of characteristic quantities of the relaxation time spectrum such as the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference the relaxation times with regard to the maxima of a distribution function of polarization resistances, the mean value of the maxima, the center of gravity of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several defined frequency ranges and the mean value of the Integrals over one or more delimited frequency ranges with sizes from the same magazine for a single battery cell or for several series and / or parallel connections

Battery cells; Comparison of characteristic quantities of the relaxation time spectrum such as the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference the relaxation times with regard to the maxima of a distribution function of polarization resistances, the mean value of the maxima, the center of gravity of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several defined frequency ranges and the mean value of the Integrals over one or more delimited frequency ranges with quantities from at least one previous time step for an individual battery cell or for several battery cells connected in series and / or in parallel;

Comparison of characteristic quantities of the relaxation time spectrum such as the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference the relaxation times with regard to the maxima of a distribution function of polarization resistances, the mean value of the maxima, the center of gravity of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several defined frequency ranges and the mean value of the Integrals over one or more delimited frequency ranges with quantities from at least one previously carried out measurement for a single battery cell or for several battery cells connected in series and / or in parallel;

Comparison of statistical variables of the relaxation time spectrum, in particular mean value and scatter measures for the parameter, value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima Distribution of the maxima in the

Spectrum, the difference in relaxation times with respect to the maxima of the distribution function of the polarization resistances, the mean value of the maxima, the center of gravity of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the

Spectrum over a certain frequency range, the integral values over several delimited frequency ranges and the mean value of the integrals over one or more delimited frequency ranges with quantities from the same magazine for a single battery cell or for several battery cells connected in series and / or in parallel;

Comparison of statistical quantities of the relaxation time spectrum, in particular mean value and scatter measures for the parameter, value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference between the relaxation times with respect to the maxima of the distribution function of the polarization resistances, the mean value of the maxima, the focus of the spectrum, the value of the spectrum at a specific frequency, the Value of the integral of

Spectrum over a certain frequency range, the integral values over several delimited frequency ranges and the mean value of the integrals over one or more delimited frequency ranges with quantities from at least one previous time step for a single battery cell or for several series and / or parallel connections

Battery cells;

Spectrum over a certain frequency range, the integral values over several delimited frequency ranges and the mean value of the integrals over one or more delimited frequency ranges with quantities from at least one previously performed measurement for an individual battery cell or for several battery cells connected in series and / or in parallel.

The term size includes in particular the characteristic and / or statistical sizes.

The term comparison also includes a comparison of several comparisons and / or one

Compare with results from one or more comparisons.

The invention thus provides in particular that at least one of the individual

Frequency components of the current excitation signal are adaptively adapted depending on the properties and the operating state of the individual battery cell or the average operating state of a plurality of battery cells connected in series and / or in parallel, and the measurement error is thus minimized.

According to the invention, it is provided in particular that several frequencies are excited simultaneously. This can significantly reduce the measurement time, so that among most

Operating conditions can be approximated from time invariance. This

The multi-frequency method is designed in particular in such a way that a continuous measurement of the instantaneous impedance of one or more battery cells is achieved by means of an excitation unit and further downstream units. For this purpose, a previously defined form of the Current excitation signal (excitation current form), which may contain, for example, 2 to 100, preferably 5 to 50, for example 20 suitably distributed, for example linearly or preferably logarithmically distributed frequency components in a range of preferably 1 Hz to 10 kHz, continuously continued on the battery cell or a plurality of Battery cells stamped and, in particular, adaptively changed.

The temporal course of the current excitation signal to be constructed can be realized by superimposing any periodic oscillations, for example sine oscillations,

Cosine vibrations, rectangular vibrations and / or triangular sequences. This is explained below using an example of a sinusoidal oscillation, this mutatis mutandis also applicable to the other forms of oscillation, in particular cosine oscillations.

Sine vibrations and cosine vibrations are advantageous in order to enable a fast Fourier transformation, in particular a Goertzel transformation, during the evaluation.

The current excitation signal is described below using N (natural number) superimposed sine waves as an example.

Here a _{k is} the amplitude, f _k the frequency and phase position Fi _<of the k-th component of the

Excitation signal. The parameter b is an additional scaling factor for that from N

Frequency components existing sum signal. The choice of these parameters depends on the operating state of the battery system. Typical real excitation units deliver a limited total current amplitude, so that depending on the choice of the individual parameters a and F of the N components, scalings are advantageously carried out, since these also have an influence on the amplitude of the overall signal. If possible, the parameter F is selected such that, as is generally preferred, there is no constructive superposition of the vibrations. The number N is chosen, in particular, depending on which interference influences are present, since the resulting amplitude of the measurement signal per frequency point depends on this. In operating states with few faults (e.g. in the idle state of the battery system), for example, several

Frequency components excited to get a more complete spectrum. In operating states with many faults (e.g. in dynamic operation), for example, fewer

Frequency components are excited in order to quickly obtain particularly relevant sections of the impedance spectrum. The choice of frequency points (parameter f) depends in particular on which effects are to be observed in the spectrum. Effects on the cathode or anode side are typically observed at different frequencies, as are effects

For example, concern the SEI (Solid Electrolyte Interphase) of a lithium-ion battery cell. The invention further relates to an excitation unit for carrying out the method, comprising a storage capacitor, a bidirectional power electronic converter and a filter, with a means for charging the storage capacitor before the start of a measurement, the storage capacitor supplying energy by means of a mean-value free excitation signal by means of the bidirectional, power electronic converter to be imprinted on one or more battery cells so that energy between the storage capacitor and the battery can be cyclically shifted. In particular, the unit also has control means for controlling the implementation of the method and / or a computing unit for calculating in particular the at least one impedance, evaluation size and / or diagnostic size and / or for carrying out the comparison and / or the adaptive control.

A possible embodiment of an excitation unit consists of a storage capacitor, a bidirectional power electronic converter and a filter. The included one

The storage capacitor must be charged before the start of the measurement, and supplies the necessary energy for a mean-free current excitation signal by means of the bidirectional,

power electronic transducer on one or more battery cells. Energy is thus cyclically shifted between the storage capacitor and the battery. This means that only the losses in the excitation unit have to be compensated. This can be done either by permanently superimposing a small DC component, which means that

Excitation signal construction, however, is no longer completely free of mean values. However, this may be tolerable if the battery is in an operating state that allows it anyway.

Alternatively, the storage capacitor can also be recharged cyclically. This design of an excitation unit also offers the advantage that the battery is technically overcharged

is excluded.

In applications such as electric vehicles, there are often changing ones

Operating and environmental conditions, such as very different temperatures, inhomogeneous charge states or aging of the battery cells or electrical faults. Therefore, the excitation unit and the characteristics of the excitation current for the impedance measurement should be very robust against these influences. The latter can be achieved by adaptively procuring the excitation signal. This concerns the frequencies contained, relative phase positions and / or amplitudes of the individual frequency components and the amplitude scaling of the superimposed overall signal.

By dynamically adapting the frequency components it contains, the influence of

Frequency ranges in which transient interference is present are minimized. This also requires a dynamic adjustment of the relative phase positions of the frequency components in order to obtain an overall signal with the lowest possible crest factor. This reduces the necessary peak power of the excitation circuit and minimizes the disturbing influence of non-linearities of the test object (the Battery cell). Simultaneously or independently of this, the amplitude of the individual frequency components is also adapted to the device under test and the external conditions. Since the magnitude of the impedance is frequency-dependent, the excitation amplitude is dynamically reduced at frequencies with magnitudes greater in magnitude in order to achieve a dynamically predetermined setpoint for the signal-to-noise ratio of the measurement signal. This reduces both the necessary excitation power and the influence of non-linearities. Similarly, the total amplitude of the excitation signal is scaled in the time domain. These functions are implemented within a control unit that generates the necessary signal for the excitation unit.

The response measurement signal (measurement signal) obtained is then preferably sampled by a measurement unit within a sliding evaluation window, the width of which corresponds to the period duration of the lowest frequency contained, or an integral multiple thereof. This can be, for example, a simple rectangular window, or it can also be a more specific variant that carries out a weighting of the individual sampling values.

A preferred embodiment of the invention is characterized in that an adaptive

Compensation of charging and discharging currents in the response measurement signal takes place, also called drift correction, which occur due to real battery operation, such as while an electric vehicle is traveling. An adaptive compensation of charge and discharge currents in the

Response measurement signal in particular takes place by comparing the voltage of the

Response measurement signal at a time t ₀ and a time t = t ₀ + NT, where T is the

Period of the smallest frequency component, and N is an integer.

The second point in time is shifted from the first point in time by the period of the smallest frequency component contained, or a multiple thereof. An interpolation can then be used to obtain a correction signal from the values obtained, which correction signal is applied to the values of the response measurement signal recorded within the measurement window in order to compensate for the voltage change in the measurement signal caused by the superimposed charging or discharging current.

Another preferred embodiment of the invention is characterized in that

Instantaneous impedances of one or more frequencies can be determined from the current excitation signal and voltage measurement signal within a measurement time window. The consequence of

Impedances are called the impedance spectrum.

A preferred embodiment of the invention is characterized in that the

Impedance spectrum or the impedance spectra using a Kramers-Kronig relation and / or further parameters on their quality characteristics, in particular consistency of at least one of the following parameters:

Signal-to-noise ratio Amplitude of the response measurement signal Linearity of the impedance spectrum Time invariance of the response measurement signal is checked and the current excitation signal is adjusted as a function thereof

The impedance values of n battery cells at k frequencies for a point in time can be in

Represent matrix form:

A - o) ^-

The recorded impedance spectra are used to determine the battery status

advantageously further processed in the diagnostic unit.

The recorded impedance, in particular the impedance spectrum, is processed in particular by statistical methods and / or by evaluating deterministic variables.

When determining the battery state, the variables real part, imaginary part, amount and / or phase angle for one or more of the excited frequencies, for example, are used for diagnosis as follows from the impedance spectrum, in particular of a first battery cell.

For example, the real part of a battery cell C1 is here at any excited frequency, referred to below as f _k , in the time window T Re (t = G)) is used. In particular, it is assumed that a previous consistency check has evaluated all impedance spectra used for the subsequent evaluation as valid.

Step 1: Form the parameter absolute change k _{t of} the current value of the real part of the battery cell C1 to be diagnosed with the value from the previous measurement time window T - 1:

Step 2: Determine a measure of scatter, for example the standard deviation s of = G)) or of k _t within the past 10 measurement time windows. A measurement time window is in particular the one used to carry out step f once.

Criterion 1: If {k ^ Ci, f _k , T) \> 2s is fulfilled, then a first criterion for the detection of a battery state change is fulfilled. Step 3: Determine the mean value M _tot and a scatter measure, for example standard deviation o _tot from Re (Z) at frequency f_k of all other N cells except cell C1 in the battery system in the time window T and T-1.

- Criterion 2: Re ((t = G))> M _fles (t = T) + 2a _fles (t = T) or Re ((t = G)) <

M _tot (t = T) - 2 a _tot (t = T), i.e. outside the band M _tot (t = T) + 2 a _tot (t = T), a second criterion for the detection of a change in battery status is fulfilled.

- T - 1)) <M _ges {t = T - 1) + 2o _saturated t = T - 1) or

s (t

= T - 1) - 2 o _tot (t = T— 1), so the real part was in the last

Time step within the band) M _tot (t = T) + 2 o _tot (t = T), a third criterion for recognizing a change in the battery status is fulfilled.

If at least two of the three criteria are met, a new measurement and subsequent determination of the battery condition should take place immediately. If at least two criteria are still met, a change in the battery state can be assumed to be recognized and, in particular, a warning signal WA (“battery state change detected”) can be output.

Then applies in particular

accelerating process rising real part before, and there may be a warning signal WB ("critical battery condition"). be issued. C is a factor that

typically between 1 and 1.5.

To further increase the confidence, the same method can also be carried out in parallel with one or more further excited frequency points f, and a change in the battery state can be assumed to be recognized and / or a warning signal WA or WB can only be output when the criteria for a predetermined plurality or are satisfied for all frequency points.

In statistical processing, statistical properties of the recorded spectra are calculated across all battery cells at the respective frequencies. This gives a measure of the inhomogeneity of the battery pack and the share of the individual battery in the inhomogeneity.

A preferred embodiment of the invention is characterized in that mean values and / or scattering dimensions are formed from the impedances to different frequencies and different battery cells that are simultaneously excited and detected in the respective measurement time window.

The mean is preferably the arithmetic, geometric and / or quadratic mean. The mean value can also be median or the central value. The measure of scatter around the arithmetic mean is preferably the variation, the variance, the standard deviation and / or the mean absolute deviation.

The measure of variation around the median is preferably the quartile distance, the interquartile distance, the mean absolute deviation with respect to the median and / or the median of the absolute deviations.

The measure of scatter around the geometric mean is preferably the geometric one

Standard deviation.

There are the characteristic quantities of the impedance, preferably real part, imaginary part, amount and phase with the sizes from the current time step for at least one individual, further battery cell and / or with the sizes of the same cell from another measurement time window and / or among several serial and / or compared battery cells of the same type connected in parallel. Depending on the scatter dimensions, threshold values can be defined here in order to detect changes in the battery state, for example undesired electrochemical processes such as electrolyte and SEI decomposition and gas formation in the battery cells. At

Exceeding a certain temperature results in irreversible changes in the

Impedance spectrum due to these phenomena. If this occurs in operation in individual battery cells of the battery pack, this can be recognized by comparison with battery cells that are not affected. Alternatively, a spectrum artificially determined from the mean or median of the impedance values can also be used for this. For this purpose, the average or median is formed over each frequency point of the impedance spectra of two or more battery cells. The consequence of this calculated

Average or median impedance values again result in an impedance spectrum. From this one or more diagnostic variable (s) can be derived, which can then be used for comparison.

A statement about the speed of the change in the spectrum can be made by differentiation over time using the spectra recorded in previous time steps (trend analysis). If rapid changes can be identified here, this can be used as an additional warning sign. Since this requires spectra to be stored digitally, a system is advantageous that thins the stored spectra over time in order to save storage space. To identify slower processes, such as cell aging, a comparison with spectra determined in the laboratory is possible in order to draw conclusions about the type of aging phenomena that occurs. Here, too, a trend analysis over larger time periods makes sense.

Another preferred embodiment of the invention is characterized in that the

Transmission unit transmits the warning signals to a control device which, based on the warning signals, warns the user and / or initiates countermeasures. Examples of possible countermeasures are the reduction of the electrical load on the battery cell, cooling and / or reinforcement of the cooling of the battery cell or the switching off of the battery cell, in particular in order to prevent the risk of its impairment or destruction.

Another preferred embodiment of the invention is characterized in that the

characteristic quantities of the impedance, preferably real part, imaginary part, amount and phase, are compared with quantities from the current time step for a single battery cell or for a plurality of battery cells connected in series and / or in parallel. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (battery state change detected) is transferred to the transmission unit. If the difference between the quantities exceeds the further threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

Another preferred embodiment of the invention is characterized in that the

characteristic quantities of the impedance, preferably real part, imaginary part, amount and phase, are compared with quantities from at least one previous time step for a single battery cell or for a plurality of battery cells connected in series and / or in parallel. This comparison can be done, for example, by simply forming differences, or by differentiating according to time. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

Difference between the sizes exceeds the further threshold value SB, a warning signal WB (safety-critical battery condition) is transferred to the transmission unit.

Another preferred embodiment of the invention is characterized in that the

characteristic quantities of the impedance, preferably real part, imaginary part, amount and phase, are compared with quantities from at least one previously carried out measurement for a single battery cell or for a plurality of battery cells connected in series and / or in parallel. This comparison can be done, for example, by simple difference formation, or by differentiation according to time. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

Difference between the characteristic quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery condition) is transmitted to the transmission unit.

Another preferred embodiment of the invention is characterized in that the

statistical values of the impedance: preferably mean and scatter measurements for the

Parameters real part, imaginary part, amount and phase with quantities from the current time step for a single battery cell or for several series and / or parallel battery cells is compared. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

Difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery condition) is transferred to the transmission unit.

Another preferred embodiment of the invention is characterized in that the

Parameters real part, imaginary part, amount and phase are compared with quantities from at least one previous time step for an individual battery cell or for several battery cells connected in series and / or in parallel. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (battery state change detected) is transferred to the transmission unit. If the difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

Another preferred embodiment of the invention is characterized in that the

Statistical values of the impedance: mean value and scatter measurements for the parameters real part, imaginary part, amount and phase are compared with values from at least one previously performed measurement for a single battery cell or for several battery cells connected in series and / or in parallel. If the difference between the quantities exceeds the threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

A further preferred embodiment of the invention is characterized in that the result of the comparison of the characteristic quantities of the impedance preferably real part, imaginary part, amount and phase with the result of the comparison from at least one previous time step for a single battery cell or for several series and / or battery cells connected in parallel is compared. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

A further preferred embodiment of the invention is characterized in that the result of the comparison of the characteristic quantities of the impedance is preferably real part, imaginary part, amount and phase with the result of the comparison from previously carried out measurements for a single battery cell or for several series and / or parallel battery cells is compared. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

A further preferred embodiment of the invention is characterized in that the result of the comparison of the statistical variables: preferably the mean and the scatter dimensions for the parameters real part, imaginary part, amount and phase with the result of the comparison from at least one previous time step for an individual battery cell or for several battery cells connected in series and / or in parallel are compared. If the difference between the quantities exceeds the threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (change in battery state determined) is transmitted to the transmission unit. If the difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

A further preferred embodiment of the invention is characterized in that the result of the comparison of the statistical variables: preferably the mean value and the scattering dimensions for the parameters real part, imaginary part, amount and phase with the result of the comparison from at least one previously carried out measurement for an individual battery cell or is compared for several battery cells connected in series and / or in parallel. If the difference between the quantities exceeds the threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (change in battery state determined) is transmitted to the transmission unit. If the difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

Additionally or alternatively, the distribution of the relaxation time constant is determined from the excitation and response measurement signals. The distribution is also known as the relaxation time spectrum (RZS) of a battery. The RZS is the calculation of the distribution of the relaxation time constants, also known as the "Distribution of Relaxation Times" (DRT). This enables frequency-dependent separation of the loss processes

Battery cell due to the property that the Kramers-Kronig relationships (see below)

Have validity, and thus the impedance of a battery cell can be understood as an infinite network of RC elements with different time constants.

Characteristic quantities are advantageously determined from the RZS and used for the condition diagnosis of the battery cell. Characteristic variables for the relaxation time spectrum are, for example, the value of the global maximum, frequency of the global maximum, the values of the local maxima, frequencies of the local maxima, number of local maxima, distribution of the local ones Maxima in the spectrum, the mean of the maxima, focus of the RZS, the value of the RZS at a certain frequency, the value of the integral of the RZS over a certain frequency range, the values of the integrals over several defined frequency ranges of the RZS, the mean value of the integrals over one or several defined frequency ranges. A diagnosis can then be carried out analogously to the method presented for impedance spectra.

The relaxation time spectrum (RZS) from the current time step for an individual battery cell or for a plurality of battery cells connected in series and / or in parallel is advantageously compared.

Depending on the dimensions of the scatter, threshold values can be defined here in order to detect undesired electrochemical processes (such as electrolyte and SEI decomposition and gas formation in the battery cells, and a change in the battery state can be detected if the temperature is exceeded. If a certain temperature is exceeded, irreversible changes occur in the

Relaxation time spectrum through these phenomena. If this occurs in operation in individual battery cells of the battery pack, this can be recognized by comparison with battery cells that are not affected and a change in the battery state can be detected when predetermined deviations are exceeded. Alternatively, an artificial relaxation time spectrum from the mean or median can be used as characteristic quantities.

The state of the individual battery and / or the overall state of the battery system is diagnosed from the analysis of the relaxation time spectra of an individual battery and the comparison of the relaxation time spectra with one or more additional batteries in the battery system. The diagnostic unit preferably carries out the diagnostic method step.

The polarization contributions of n battery cells at k frequencies for a point in time can be represented in matrix form:

ß (t ₀ )

For the diagnosis of the operating and safety parameters mentioned at the outset, further processing of the recorded relaxation time spectra is advantageous in order to use them

Generate diagnostic variables. This is carried out by the diagnostic unit. This processes the recorded spectra in several ways:

Statistical properties of the recorded spectra over all battery cells at the respective frequencies are calculated, such as median, mean or scatter measures like that

Interquartile range or standard deviation. This gives a measure of the inhomogeneity of the battery state of the battery pack.

Another preferred embodiment of the invention is characterized in that the

characteristic sizes of the relaxation time spectrum: preferably value of the global Maximums, frequency of the global maximum, the values of the local maxima, frequency of the local maxima, number of local maxima, distribution of the local maxima in the spectrum, the mean value of the maxima, center of gravity of the signal, the value of the spectrum at a specific frequency, the value the integral of the spectrum over a certain frequency range, the integral values over several delimited frequency ranges, the mean value of the integrals over one or more delimited frequency ranges with quantities from the current time step for a single one

Battery cell or for several series and / or parallel battery cells is compared. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (battery state change detected) is transferred to the transmission unit. If the difference between the quantities exceeds the further threshold value SB, a warning signal WB (safety-critical

Battery status) to the transmission unit.

Another preferred embodiment of the invention is characterized in that the

characteristic sizes of the relaxation time spectrum: preferably value of the global maximum, frequency of the global maximum, the values of the local maxima, frequency of the local maxima, number of local maxima, distribution of the local maxima in the spectrum, the mean value of the maxima, focus of the signal, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a certain frequency range, the integral values over several delimited frequency ranges, the mean value of the integrals over one or more delimited frequency ranges with quantities from at least one previous time step for a single battery cell or for several series and / or battery cells connected in parallel is compared. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and in particular a warning signal WA

(Battery state change detected) transferred to the transmission unit. If the

Another preferred embodiment of the invention is characterized in that the

characteristic sizes of the relaxation time spectrum: preferably value of the global maximum, frequency of the global maximum, the values of the local maxima, frequency of the local maxima, number of local maxima, distribution of the local maxima in the spectrum, the mean value of the maxima, focus of the signal, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a certain frequency range, the integral values over several delimited frequency ranges, the mean value of the integrals over one or more delimited frequency ranges with quantities from at least one previously carried out

Measurement, for example in the laboratory environment, is compared for a single battery cell or for several battery cells connected in series and / or in parallel. If the difference between the quantities exceeds a threshold value SA, a change in the battery state is recognized and In particular, a warning signal WA (battery state change detected) is transferred to the transmission unit. If the difference between the characteristic quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery condition) is sent to the

Transfer unit handed over.

A further preferred embodiment of the invention is characterized in that the result of the comparison of the statistical variables of the relaxation time spectrum: preferably mean value and scatter measures for the value of the global maximum, frequency of the global maximum, the values of the local maxima, frequency of the local maxima, number the local maxima, distribution of the local maxima in the spectrum, the mean value of the maxima, focus of the signal, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several defined frequency ranges, the mean value of the Integrals over one or more delimited frequency ranges are compared with variables from the current time step for a single battery cell or for several battery cells connected in series and / or in parallel. If the difference between the sizes

If the threshold value exceeds SA, a change in the battery state is recognized and, in particular, a warning signal WA (battery state change detected) is transferred to the transmission unit.

If the difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

A further preferred embodiment of the invention is characterized in that the result of the comparison of the statistical variables of the relaxation time spectrum: preferably mean value and scatter measures for the value of the global maximum, frequency of the global maximum, the values of the local maxima, frequency of the local maxima, number the local maxima, distribution of the local maxima in the spectrum, the mean value of the maxima, focus of the signal, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several defined frequency ranges, the mean value of the Integrals over one or more delimited frequency ranges are compared with quantities from at least one previous time step for an individual battery cell or for several battery cells connected in series and / or in parallel. If the difference between the quantities exceeds the threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (change in battery state determined) is transmitted to the transmission unit. If the difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

A further preferred embodiment of the invention is characterized in that the result of the comparison of the statistical variables of the relaxation time spectrum: preferably mean value and scatter measures for the value of the global maximum, frequency of the global maximum, the values of the local maxima, frequency of the local maxima, number the local maxima, distribution of the local maxima in the spectrum, the mean of the maxima, focus of the signal, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over a specific frequency range, the integral values over several delimited frequency ranges, the mean value of the integrals over one or more delimited Frequency ranges with quantities from at least one previously carried out measurement for an individual battery cell or for a plurality of battery cells connected in series and / or in parallel are compared. If the difference between the quantities exceeds the threshold value SA, a change in the battery state is recognized and, in particular, a warning signal WA (change in battery state determined) is transmitted to the transmission unit. If the difference between the quantities exceeds the threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

A statement about the speed of the change in the relaxation time spectrum can be made by differentiation over time using the spectra recorded in previous time steps (trend analysis). If rapid changes can be identified here, this can be used as an additional warning sign. Since this requires spectra to be stored digitally, a system is advantageous that thins the stored spectra over time in order to save storage space. To identify slower processes, such as cell aging, a comparison with spectra determined in the laboratory is possible in order to draw conclusions about the type of aging phenomena that occurs. Here, too, a trend analysis over larger time periods makes sense.

It is preferably provided in the invention that at least one of the individual

Frequency components of the current excitation signal are adaptively adapted depending on the properties and the operating state of the individual battery cell or the average operating state of a plurality of battery cells connected in series and / or in parallel, and the measurement error is thus minimized. The current excitation signal consists of at least two periodic signals with mutually different frequencies and is designed such that it is free of averages over at least one period of the smallest frequency contained, so that a first response measurement signal of the battery cell is determined and evaluated that at least one of the parameters amplitude, frequency and the relative phase position of at least one component of the current excitation signal is changed as a function of the first response measurement signal, so that a measurement error is minimized, a further response measurement signal is determined and evaluated and that the evaluated value is used as an evaluation variable.

The invention further relates to an excitation unit for carrying out the method comprising a storage capacitor, a bidirectional power electronic converter and optionally a filter, with a means for charging the storage capacitor before the start of a measurement, the storage capacitor supplying energy in order to generate a mean-free current excitation signal by means of the bidirectional, power electronic Imprint converter on one or more battery cells so that energy between the storage capacitor and battery can be cyclically shifted. The invention further relates to a device for carrying out the method comprising a control unit, an excitation unit, a measuring unit, an evaluation unit, a

Control unit, a diagnostic unit and a transmission unit.

Advantageously, the device is characterized in that the control unit is set up to generate an excitation signal construction with suitable parameters, in particular excitation frequencies, excitation amplitudes, phase positions and a total amplitude scaling (scaling factor), that the excitation unit is further equipped so that it is adaptive

Procurement of the excitation signal construction enables, and in particular is set up to adaptively adapt at least one of the parameters of the excitation signal construction.

The individual components, in particular the control unit and the excitation unit, can be combined both individually and on a circuit board or combined in a single unit. The individual components can also be implemented as software units.

An advantageous further development of the device is characterized in that the measuring unit (a means for recording the impedance spectrum) is set up for simultaneous detection and recording of the voltage measurement signal in a previously defined measurement time window by one or more simultaneously excited battery cells.

A useful development of the device is characterized in that the

Control unit for generating an excitation signal construction of a characteristic multifrequency periodic, continuous signal defined in a control unit is set up in one or at the same time a plurality of battery cells connected in series and / or in parallel.

An advantageous development of the device is characterized in that it has a

Contains control unit that is set up in such a way that it uses the recorded voltage values to determine adaptive correction parameters for the amplitudes, frequencies and relative phase positions for the next measurement time window and transmit them to a control unit.

An expedient development of the device, with an excitation unit for applying a current excitation signal; with a means for recording an impedance spectrum of the battery cell; with a means for determining an evaluation variable based on the measured

Impedance spectrum is characterized in that the excitation unit (the means for applying a current excitation signal) is designed in such a way that the current excitation signal consists of at least two periodic signals with mutually different frequencies, that the current excitation signal can be applied in such a way that it is free of mean values over at least one period of the smallest frequency contained, that the means for recording the impedance spectrum of the battery cell is designed such that a first response measurement signal of the battery can be determined and evaluated and that at least one of the The parameter amplitude, frequency and relative phase position of at least one of the signals can be changed as a function of the first response measurement signal, so that a measurement error is minimized, a further response measurement signal is determined and evaluated, that the evaluated value can be used as an evaluation variable and that the determination the battery state of the battery cell is carried out on the basis of a comparison of the diagnostic variable with a reference value and / or with at least one further diagnostic variable.

Description of preferred embodiments of the invention with reference to the drawings

Brief description of the pictures

These and other aspects of the invention are shown in detail in the figures as follows:

Fig.1: A group of impedance spectra before, during and after the onset of

Electrolyte decomposition and gas formation

Fig.2: Device for performing the method according to the invention.

Fig.3: Sequence of a preferred method

4: Embodiment for the excitation unit, consisting of a storage capacitor, a bidirectional power electronic converter and optionally one or more filter elements

Fig. 5: A bevy of relaxation time spectra before and during the onset of

Electrolyte decomposition and gas formation

Fig. 6: A relaxation time spectrum after the onset of electrolyte decomposition and gas formation, in comparison with the relaxation time spectra before and during the onset of

Electrolyte decomposition and gas formation.

A possible embodiment of an advantageous device for performing the invention

The method is shown in Fig. 2. It consists of a control unit 10, a means for

Generation of an external trigger signal 15, an excitation unit 20, a measuring unit 30, one

Evaluation unit 40, a control unit 50, a diagnostic unit 60 and one

Transmission unit 70. The means for generating the external trigger signal 15 and the trigger signal 15 are optional and make it possible to start a new measurement cycle as a function of external events.

The external trigger signal 15 is preferably transmitted via a data bus (not shown).

With this device according to the invention, methods according to the invention can be carried out.

A preferred implementation of the method according to the invention is explained below with reference to FIG. 3. This method is expediently carried out using the device shown in FIG. 2.

An excitation signal construction 100 takes place with suitable parameters, in particular

Excitation frequencies, excitation amplitudes, phase positions and a total amplitude scaling (scaling factor).

A preferred embodiment provides for the excitation frequencies to be selected such that they do not have higher harmonics, in particular 2nd and / or 3rd Harmonics, other excited

Frequencies are overlapping.

Since there are frequently changing operating and ambient conditions in applications such as electric vehicles, such as very different temperatures, inhomogeneous charge states or aging of the battery cells or electrical disturbances, the excitation unit and the characteristics of the excitation current for the impedance measurement must be very robust these influences. The latter can be achieved by adaptively procuring the excitation signal. This applies to the contained frequencies, relative phase positions and amplitudes of the individual frequency components and the amplitude scaling of the superimposed overall signal.

By dynamically adapting the frequency components it contains, the influence of

Frequency ranges in which transient interference is present are minimized. This also requires a dynamic adjustment of the relative phase positions of the frequency components in order to obtain an overall signal with the lowest possible crest factor. This reduces the necessary peak power of the excitation circuit and minimizes the disturbing influence of non-linearities of the test object (the battery cell). At the same time or independently, the amplitude of the individual

Frequency components adapted to the device under test and the external conditions. Since the magnitude of the impedance is frequency-dependent, the method can be used to reduce the excitation amplitude at frequencies with larger magnitudes by a dynamically predetermined one

To reach the target value for the signal-to-noise ratio of the measurement signal. This also reduces the necessary excitation power and the influence of non-linearities. In a similar development, the method according to the invention can be used in an advantageous development of the invention

Scale the total amplitude of the excitation signal in the time domain. These functions are within implemented a control unit that generates the necessary control signal for the excitation unit. In this case, from a control unit, the in the previous step through the

Evaluation unit determined impedance values the respective setpoints for the next

Measurement run determined and transmitted to the control unit.

After the excitation construction 100 has been carried out in the control unit 10, a current excitation signal is generated 1 10 by the excitation unit 20. The excitation unit 20 is also referred to as a driver.

A possible embodiment of an excitation unit 20 according to the invention is shown in FIG. 4. It consists of a storage capacitor, a bidirectional power electronic converter and a filter. The storage capacitor contained must be charged before the start of the measurement and supplies the energy required to apply a mean-free current excitation signal to one or more battery cells by means of the bidirectional, power electronic converter. Energy is thus cyclically shifted between the storage capacitor and the battery. This means that only the losses in the excitation unit have to be compensated. This can be done either by permanently superimposing a small DC component, which means that

Excitation signal construction, however, is no longer completely free of mean values. Battery conditions that tolerate such excitation are possible depending on the application. Alternatively, the

Storage capacitor can also be recharged cyclically. This design of an excitation unit also offers the advantage that overcharging of the battery can be technically excluded. The excitation unit 20 can be any excitation signal constructions 100 for the

Generate diagnosis 185 according to the invention. It is also possible to use the invention

Combine excitation signal construction 100 with any diagnosis 185. A voltage response measurement 120 is carried out in a subsequent method step. Thereafter, a drift correction 130 and a transformation 140 in the frequency domain

(Frequency domain transformation). The frequency domain transformation 140 can advantageously be carried out with the so-called Görtzel algorithm, the one

Special form of the discrete Fourier transformation (DFT) is. Compared to the better known "Fast Fourier Transform" (FFT), this has the advantage of higher computing time efficiency, for the here

typically present case that only comparatively little discrete spectral components, typically exactly those which were also contained in the excitation signal, are required. Alternatively, a wavelet transformation is also advantageous if non-periodic or periodic non-trigonometric excitation functions are used. Also at

Voltage response measurements in which the measurement signal is windowed, a wavelet transformation is advantageous.

An impedance spectrum determination 150 and an optional relaxation time determination 160 are preferably carried out simultaneously. It is also possible to determine 160 of the Omitting the relaxation time spectrum or performing the relaxation time determination following a consistency check 170.

The consistency check 170 enables adaptive control 180 of at least one of the parameters of the excitation signal construction 100. The consistency check serves to assess the validity of the recorded impedance spectra. A common basic requirement for

Electrochemical impedance spectroscopy is that the behavior of the system under investigation can be described as linear and time-invariant. These criteria are met by

Consistency check.

The consistency check 170 is advantageously carried out in several ways. In the event that a

Frequency spectrum of the measurement signal is present, this is examined for harmonics contained in the excitation signal. If harmonics are detected, this is an indication of non-linear behavior. Subsequently, the excitation signal is modified by the adaptive control 180 and the adapted excitation signal construction 100. Harmonics are no longer recognized.

Further advantageous criteria for the consistency check 170 and assessment of time-invariant behavior are the Kramers-Kronig relations (set out, for example, in the following publication: M. Schönleber, D. Klotz and E. Ivers-Tiffee, A Method for Improving the Robustness of linear Kramers-Kronig Validity Tests, Electrochimica Acta 131, pp. 20-27 (2014)), which relate the real and imaginary parts of the recorded impedance spectra, and check for consistency. If during the consistency check 186 an excessive deviation between measured real parts or measured imaginary parts of the impedance and impedance or impedance constructed using Kramers-Kronig relations is found, a violation of the time invariance criterion is recognized and the measurement is recognized as inconsistent and rejected. The adaptive control 180 and the adapted excitation signal construction 100 construct a new excitation signal and repeat the measurement.

During a dynamic load of the battery with a superimposed load current or

Charging current, the consistency check 170 of the impedance measurement is carried out additionally or as a substitute with analysis of the current strength. If the load current or charging current changes more than a tolerance band D during the measurement, the measurement is recognized as inconsistent because the state of the battery has changed to such an extent that there is no longer any time invariance. Depending on the current strength of the superimposed current, the adaptive control 180 and the adapted

Excitation signal construction 100 eliminates the low frequency components and reduces the measurement duration to such an extent that time-invariant behavior is again present in the next measurement step and the necessary consistency is restored. Another feature of the consistency check 170 in the evaluation unit 40 is a fault analysis of the measurement signal. For this purpose, the signal-to-noise ratio of the measurement signal is evaluated, for example. If a high spectral noise power density is close to the excited one

Frequencies is present, this is an indication of interference. The at least one frequency or amplitude of the excitation signal is changed by the adaptive control 180 and the adapted excitation signal construction 100 in order to increase the signal-to-noise ratio.

The adaptive control 180 processes the results of the consistency check further and defines the requirements for a new excitation signal. For example, the amplitude of the current excitation signal must be reduced in the next time step if non-linearities are determined. If strong interference is found in certain frequency bands, the excitation frequencies are varied. If a violation of the time invariance criterion is detected, the measurement duration is adaptively shortened according to the invention by omitting low frequencies and / or recording them incompletely.

The evaluation unit 40 preferably carries out the method steps 130 to 170. The measurement data is then used for diagnosis 185. Finally, another

Reduction of the required memory space 186 by thinning out spectra over time.

1 shows a family of impedance spectra of an exemplary lithium-ion battery at different temperatures. The representation takes the form of a so-called Nyquist diagram, in which the real part of the impedance is plotted on the X axis and the negative imaginary part of the complex impedance is plotted on the Y axis.

The impedance spectra 300-320, corresponding to a temperature range of 40 ° C.-60 ° C., have a normal behavior in that the charge transfer process, which is typically recognizable as a compressed semicircle, of one or both electrodes becomes smaller, which is one

Corresponding reduction in the associated losses. The at 70 ° C, 80 ° C and 90 ° C

Recorded impedance spectra 330 - 350 already show increasing abnormalities, which are expressed in the fact that there is an increasing shift towards larger values on the axis of the real part. This is typical of the decomposition of the electrolyte or SEI that begins in this temperature range, and the subsequent gas formation. With further increasing temperatures of 100 ° C and 110 ° C (impedance spectra 360 and 370), even more striking features come to light, which are expressed in an additional semicircle in the higher-frequency part of the spectrum in the Nyquist diagram. The impedance spectrum 380 was recorded during the subsequent cooling phase at a temperature of 80 ° C., and shows the same at that before

Temperature recorded spectrum 340 significant changes. The spectrum 390 recorded after the initial temperature of 40 ° C. has been reached again shows a permanent change in the impedance spectrum, which means severe, permanent damage to the battery cell. The analysis 185 of the state of the individual battery and / or the overall state of the battery system is carried out from the analysis of the frequency spectra of an individual battery cell and the comparison of the frequency spectra with further battery cells in the battery system. The diagnostic unit 60 preferably carries out the method step diagnosis 185.

The impedance values of n battery cells at k frequencies for a point in time can be represented in matrix form:

A - o)

Further processing of the recorded impedance spectra is advantageous for the diagnosis of the operational and safety parameters mentioned at the outset in order to generate diagnostic variables therefrom. This is carried out by the diagnostic unit 60. This processes the

recorded spectra in several types:

The characteristic quantities of the impedance real part, imaginary part, amount and phase are compared with the quantities from the current time step for a single battery cell or for several battery cells connected in series and / or in parallel. Depending on the scatter dimensions, threshold values are defined in order to detect undesired electrochemical processes, such as electrolyte and SEI decomposition and gas formation in the battery cells. As already explained with reference to FIG. 1, irreversible changes in the impedance spectra 380 and 390 result from degeneration phenomena when a certain temperature is exceeded. By comparing the spectra between different battery cells of the battery pack, one can be affected

Battery cell or several affected battery cells can be identified. For example, if battery cell A has spectrum 330 and battery cell B has spectrum 350, diagnostic unit 60 is able to identify battery cell B as a defective battery. Alternatively, a spectrum artificially determined from the mean and / or median of the impedance values is also used for the comparison. Alternatively or additionally, impedance spectra from previously carried out measurements, for example under laboratory conditions, are used as a reference for the comparison.

A statement about the speed of the change in the spectrum over time is also used for the diagnosis 185. Differentiation over time also uses the spectra recorded in one or more previous time steps Diagnosis. The process is also known as trend analysis. If rapid changes are detected here, this is used in addition or as some indicator for the diagnosis 185 of a critical condition. As an alternative or in addition, statistical variables of the impedance spectra, such as mean values and / or scatter measures, are used for the trend analysis. As an alternative or in addition, trends for the diagnosis 185 of the battery and / or the battery system become trends from previously

performed measurements, for example under laboratory conditions, used as a reference.

As an alternative or in addition there is a for the detection of slow processes, such as cell aging

Comparison with impedance spectra determined in the laboratory is provided in order to draw conclusions about the type of aging phenomena occurring. Again, a trend analysis is about larger ones

Time periods make sense.

The determination 160 of the relaxation time is the calculation of the distribution of the relaxation time constants, also known as the “Distribution of Relaxation Times” (DRT). The distribution is also known as the relaxation time spectrum (RZS) of a battery. This enables frequency-dependent separation of the loss processes of the battery cell due to the property that the Kramers-Kronig relationships are valid, and thus the impedance of a battery cell can be understood as an infinite network of RC elements with different time constants.

5 shows a family of relaxation time spectra of an exemplary lithium-ion battery at different temperatures. The frequency, which is inversely proportional to the relaxation time, is plotted on the X axis. The amount of the distribution function of the

Polarization resistances are shown on the Y axis, which are proportional to the loss processes in the battery cell.

The relaxation time spectra 400 and 402, corresponding to a temperature range of 40 ° C - 50 ° C, have normal behavior in that three typical ones in this example

Relaxation processes 460, 465 and 470 are recognizable. When the temperature rises to 60 ° C 405 and 80 ° C 410, when the first safety-critical processes start, process A 460 and process C 470 decrease significantly. These abnormalities indicate the first safety-related processes in the battery cell, such as the decomposition of the electrolyte or SEI, and the subsequent gas formation.

If the temperature continues to rise to 100 ° C 415, even more striking features come to light, which can be seen in the absolute dominance of the relaxation process A 460 in spectrum 415. 6 shows the relaxation time spectrum 420 of a battery at a temperature of 40 ° C. after cooling of 110 ° C. After reaching the initial temperature of 40 ° C again

recorded spectrum 420 shows a permanent change in the distribution of the

Relaxation time constants and the absolute dominance of the relaxation process A 460 even after the battery has cooled down, which indicates strong, permanent damage to the battery cell. From the analysis of the relaxation time spectra of an individual battery and the comparison of the relaxation time spectra with one or more additional batteries in the battery system, the diagnosis 185 of the condition of the individual battery and / or the overall condition of the battery is made

Battery system. The diagnostic unit 60 preferably carries out the method step diagnosis 185.

B (to)

For the diagnosis of the operational and safety parameters mentioned at the beginning, there is a

Further processing of the recorded relaxation time spectra is advantageous to get from it

Generate diagnostic variables. This is carried out by the diagnostic unit 60. This processes the recorded spectra in several ways:

The characteristic quantities of the values of the global maximum 430, the frequency of the global maximum 430, the values of the local maxima 435, the frequency of the local maxima 435, the number of local maxima 435, the distribution of the local maxima 435 in the spectrum, the mean value are shown the Maxima 435, the center of gravity of the signal, the value of the spectrum at one determine

Frequency 440, the value of the integral 450 of the spectrum over a certain frequency range, the integral values over several delimited frequency ranges, the mean value of the integrals over one or more delimited frequency ranges. The sizes from the current

Time steps for a single battery cell or for several battery cells connected in series and / or in parallel are compared. Depending on the measures of scatter, threshold values will be defined in order to detect undesired electrochemical processes such as electrolyte and SEI decomposition and gas formation in the battery cells. As already explained with reference to FIG. 5, irreversible changes in relaxation time spectra 420 result from degeneration phenomena when a certain temperature is exceeded. By comparing the spectra between

different battery cells of the battery pack, an affected battery cell or several affected battery cells can be identified. For example, if battery cell A has spectrum 400 and battery cell B has spectrum 415, diagnostic unit 60 is able to identify battery cell B as a defective battery cell. Alternatively, a is used for the comparison spectrum artificially determined from the mean and / or median of the relaxation time spectra can be used. Alternatively or additionally, relaxation time spectra from measurements carried out previously, for example under laboratory conditions, are used as a reference for the comparison.

A statement about the speed of the change in the relaxation time spectrum over time is also used for diagnosis 185. By differentiating over time is done under

Use of the spectra recorded in the previous time steps. The process is also known as trend analysis. If rapid changes are detected here, this is used in addition or as some indicator for the diagnosis 185 of a critical condition. As an alternative or in addition, statistical characteristic quantities of the

Relaxation time spectrum, such as mean and / or median and / or the standard deviation used for the trend analysis. Alternatively, the battery 185 and / or the

Battery system trends from previous measurements, for example under

Laboratory conditions, used for reference.

To detect slower processes, such as cell aging, a comparison with relaxation time spectra determined in the laboratory is possible in order to draw conclusions about the type of aging phenomena that occurs. Here, too, a trend analysis over larger time periods makes sense.

Since spectra have to be stored digitally for this purpose, a system is advantageous which reduces the number of stored spectra 186. The storage space for this can be reduced by thinning out spectra over time. The spectra to be deleted are selected by taking into account the operating conditions at the time the spectrum was recorded and stored, such as state of charge and temperature. If a valid spectrum is recorded at a later time under the same operating conditions as an earlier spectrum, and is due to the comparison with the spectra recorded by other battery cells

Battery cells are sufficiently certain that there is no unusual state deviation, for example the oldest spectrum stored for this state can be deleted from the buffer memory, and the new spectrum can be stored in the buffer memory with a time stamp. Through the procedure according to the invention, the storage space required for the

Spectra kept for diagnostic purposes minimized. The reduction of the data is optionally carried out by the diagnostic unit 60 or the evaluation unit 40. List of reference numbers

10 control unit

15 External trigger signal

20 excitation unit

30 measuring unit

40 evaluation unit

50 control unit

60 diagnostic unit

70 transmission unit

100 excitation signal construction

110 Generate a current excitation signal

120 voltage response measurement

130 drift correction

140 Frequency domain transformation.

150 Impedance spectrum determination

160 Determination of the relaxation time spectrum

170 consistency check

180 Adaptive control

185 diagnosis

186 Space reduction

190 storage capacitor

200 bidirectional power electronic converters

210 filters

220 battery cell; Battery string

300 impedance spectrum of a battery cell at a temperature of 40 ° C 310 impedance spectrum of a battery cell at a temperature of 50 ° C Impedance spectrum of a battery cell at a temperature of 60 ° C

Impedance spectrum of a battery cell at a temperature of 70 ° C

Impedance spectrum of a battery cell at a temperature of 80 ° C

Impedance spectrum of a battery cell at a temperature of 90 ° C

Impedance spectrum of a battery cell at a temperature of 100 ° C

Impedance spectrum of a battery cell at a temperature of 110 ° C

Impedance spectrum of a battery cell at a temperature of 80 ° C after heating up to 110 ° C

Impedance spectrum of a battery cell at a temperature of 40 ° C after heating up to 110 ° C

Relaxation time spectrum of a battery at a temperature of 40 ° C

Relaxation time spectrum of a battery at a temperature of 50 ° C

Relaxation time spectrum of a battery at a temperature of 60 ° C

Relaxation time spectrum of a battery at a temperature of 80 ° C

Relaxation time spectrum of a battery at a temperature of 100 ° C

Relaxation time spectrum of a battery at a temperature of 40 ° C after warming up to 1 10 ° C

Global maxima of a relaxation time spectrum

Local maximum of a relaxation time spectrum

Value of the relaxation time constant at 10 Hz

Integral of the relaxation time spectrum in the frequency range from 100 Hz to 1 kHz relaxation process A

Relaxation process B

Relaxation process C

Claims

1. Method for generating an adapted excitation signal for determining a

Battery state and / or a change in a battery state of at least one, in particular one, battery cell by means of impedance measurement comprising the steps a. Applying a current excitation signal to the at least one battery cell, the current excitation signal being superimposed in particular from at least two, in particular at least 5, preferably at least 15 and / or at most 25

Components, which each represent periodic signals with mutually different frequencies, exist, the current excitation signal being applied such that the mean value of the

Current excitation signal over at least one period of the smallest frequency contained is at most 20% of the peak value, in particular the current excitation signal is free of mean values over at least one period of the smallest frequency contained, b. Detecting a response measurement signal of the at least one battery cell, wherein

in particular by means of the first response measurement signal, at least one impedance, in particular an impedance spectrum, of the battery cell is determined and c. Determining at least one evaluation variable from the response measurement signal and / or the at least one impedance, characterized in that d. at least one of the parameters amplitude, in particular around 0.01 to 10 A,

Frequency, in particular around 1 Hz to 10 kHz, and relative phase position, in particular around 0 to 360 °, of at least one component of the current excitation signal in

Dependency on the first response measurement signal and / or the evaluation variable is changed, in particular so that a measurement error is minimized, e. Generating a changed current excitation signal as an adapted excitation signal, in particular in such a way that the mean value of the changed current excitation signal over at least one period of the lowest frequency contained is at most 20% of the peak value, in particular the changed current excitation signal over a period of the lowest frequency contained is mean-free

2. The method according to claim 1, wherein steps a to e are carried out several times, in particular continuously, the changed current excitation signal being applied to the at least one battery cell in such a way that the mean value of the changed

Current excitation signal over at least one period of the smallest contained frequency is at most 20% of the peak value, especially the changed one

Current excitation signal is free of mean values over at least one period of the smallest frequency contained.

3. Method for determining a battery condition and / or a change in

Battery state of at least one, in particular one, battery cell comprising the steps a. Applying a current excitation signal to the at least one battery cell, the current excitation signal consisting in particular of at least two, in particular at least 5, preferably at least 15 and / or at most 25, in particular superimposed components, each of which represents periodic signals with frequencies different from one another, the current excitation signal thus is applied that the mean value of the current excitation signal over at least one period of the lowest frequency contained is at most 20% of the peak value, in particular that

Current excitation signal is free of mean values over at least one period of the smallest frequency contained, b. Detecting a response measurement signal of the at least one battery cell, wherein

Frequency, in particular by 1 Hz to 10 kHz, and relative phase position, in particular by 0 to 360 °, of at least one component of the current excitation signal is changed as a function of the first response measurement signal and / or the evaluation variable, in particular so that a measurement error is minimized, e. that the changed current excitation signal is applied to the at least one battery cell in such a way that the mean value of the changed current excitation signal over at least one period of the lowest frequency contained is at most 20% of the peak value, in particular the changed current excitation signal over at least one period of the lowest frequency contained is mean-free and f. that a further response measurement signal of the at least one battery cell is detected, in particular at least one further impedance, in particular a further impedance spectrum, of the battery cell being determined by means of the first response measurement signal, and g. that at least one diagnostic variable is determined on the basis of the measured at least one further impedance and / or the response measurement signal and h. that the determination of the battery status and / or a battery status variable of the at least one battery cell on the basis of a comparison of the at least one diagnostic variable with at least one reference value, in particular obtained in characterization measurements, and / or with at least one further diagnostic variable, in particular the at least one battery cell and / or another Battery cell of the same type from a previous time and / or another

Battery cell of the same type from a time window, in particular with a duration such that within the time window the change in the state of charge of the at least one battery cell is less than 5% and / or the change in temperature is less than 5 K and / or the change in capacity is less than 5 % and / or of a maximum of 600 seconds around step f, is carried out, with a change in particular in the event of a significant deviation

Battery state is assumed and / or a critical battery state is assumed when the reference value is reached and / or exceeded.

4. The method according to any one of the preceding claims, characterized in that the method is carried out during the charging and / or discharging of the at least one battery cell and thereby charging and / or discharging currents, in particular due to a real battery operation, such as occur during a Driving an electric vehicle and an adaptive computational compensation of charging and / or discharging currents in the

Response measurement signal is given.

5. The method as claimed in one of the preceding claims, characterized in that the impedance spectrum is checked by means of a Kramers-Kronig relation and / or further parameters, in particular also its quality characteristics and / or quality of evaluation, in particular at least one of the following parameters:

Signal-to-noise ratio Amplitude of the response measurement signal

Linearity of the impedance spectrum, time invariance of the response measurement signal is determined and / or used.

6. The method according to any one of the preceding claims, characterized by the following steps, an excitation signal construction (100) with suitable parameters, in particular

Excitation frequencies, excitation amplitudes and phase positions, generation (110) of a current excitation signal according to the excitation signal construction by the excitation unit (20) and / or a voltage response measurement (120) and / or a drift correction (130) and / or a Fourier and / or wavelet transformation (140 ) in the frequency range, in particular an impedance spectrum determination (150) and a determination (160) of the relaxation time spectrum preferably taking place simultaneously or immediately in succession, and / or a consistency check (170) by means of an adaptive control (180) of at least one of the parameters of the excitation signal construction ( 100) is influenced.

7. The method according to any one of the preceding claims, characterized in that steps a to g are arranged on at least two, in particular at least four, battery cells, which are / are arranged in particular within a radius of 50 cm, in particular 10 cm, around one of the at least two cells , in particular of the same type, can be carried out simultaneously or sequentially within a predetermined measurement time window, the measurement time window being defined in that a change in the state of charge of the at least two battery cells is less than 5%, preferably less than 1%, a change in temperature is less than 5 K, preferably less 1 K and a change in the capacitance is less than 5%, preferably less than 1% and in particular in step f the diagnostic variables of the at least two battery cells are compared.

8. The method according to any one of the preceding claims, characterized in that steps a to g, in particular at the latest, are carried out again when the

Change in the state of charge of the battery cell is greater than 5%, or the change in temperature is greater than 5 K or the change in capacity is greater than 5%, or a time of 60 seconds, preferably 1 second or if an external trigger signal (15) is a new one Measurement requests, or that the measurement is carried out continuously with a sliding evaluation time window.

9. The method according to any one of the preceding claims, characterized in that for a renewed implementation of step a, in particular steps a to h. at least one parameter, in particular amplitude, frequency and relative phase position of the

Frequency points, of the excitation current signal, in particular a plurality or all, and / or at least one component, in particular a plurality of components or all components of the modified current excitation signal from a previous measurement, in particular carrying out steps a to 3, in particular up to h, as a starting value and / or is selected to generate the excitation current signal of step a.

10. The method according to any one of the preceding claims, characterized in that

Average values and / or scatter measurements are formed from impedances of different frequencies recorded in a respective measurement time window and / or simultaneously, in a first and one or more further battery cells of the same type.

11. The method according to any one of the preceding claims, characterized in that at least one of the characteristic quantities of the impedance such as real part, imaginary part, amount and phase of a single, first of the at least one battery cell with quantities from the same time step and / or time interval of a single second battery cell or for a plurality of further battery cells connected in series and / or in parallel, and that if a difference between the quantities exceeds a threshold value SA, a warning signal WA (battery state change detected) is transmitted to the transmission unit and that if the difference between the quantities a further threshold value SB exceeds a warning signal WB (safety-critical battery status) is transferred to the transmission unit.

12. The method according to any one of the preceding claims, characterized in that at least one of the characteristic quantities of the impedance such as real part,

Imaginary part, amount and phase with quantities from at least one previous time step and / or interval for a first battery cell or for one or more serial and / or other battery cells connected in parallel are compared, and that if a difference between the sizes exceeds a threshold value SA, a warning signal WA

(Battery state change detected) is transferred to the transmission unit and that if the difference between the quantities exceeds a further threshold value SB, a warning signal WB (safety-critical battery state) is transferred to the transmission unit.

13. The method according to any one of the preceding claims, characterized in that characteristic quantities of the impedance such as real part, imaginary part, amount and phase are compared with quantities from at least one previously carried out measurement for a single battery cell or for several battery cells connected in series and / or in parallel , and that if a difference between the sizes exceeds a threshold value SA, a warning signal WA (battery state change detected) to the

Transfer unit is transferred and that if the difference between the sizes exceeds a further threshold value SB, a warning signal WB (safety-critical battery condition) is transferred to the transfer unit.

14. The method according to any one of the preceding claims, characterized in that statistical quantities of the impedance, in particular mean value and scatter measurements for the parameters real part, imaginary part, amount and phase of a single, first battery cell with quantities from the same time step for a single, second battery cell or are compared for a plurality of further battery cells connected in series and / or in parallel, and that if a difference between the quantities exceeds a threshold value SA, a warning signal WA (change in battery state determined) is transmitted to the transmission unit and that if the difference between the quantities exceeds a further threshold value SB a warning signal WB (safety-critical battery status) is transferred to the transmission unit.

15. The method according to any one of the preceding claims, characterized in that statistical quantities of the impedance, in particular mean and scatter measurements for the parameter real part, imaginary part, amount and phase with quantities from at least one previous time step for a single battery cell or for several series and / or battery cells connected in parallel can be compared that if a difference between the sizes exceeds a threshold value SA a warning signal WA

16. The method according to any one of the preceding claims, characterized in that statistical quantities of the impedance, in particular mean value and scatter measurements for the parameters real part, imaginary part, amount and phase with quantities from at least one previously carried out measurement for a single battery cell or for several series and / or battery cells connected in parallel are compared, and that if a difference between the sizes exceeds a threshold value SA, a warning signal WA

17. The method as claimed in one of the preceding claims, characterized in that characteristic quantities of the relaxation time spectrum, such as the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, Number of maxima, the distribution of the maxima in the spectrum, the difference between the relaxation times with regard to the maxima of a distribution function of polarization resistances, the mean value of the maxima, the focus of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over one determined frequency range, the integral values over several delimited frequency ranges and the mean value of the integrals over one or more delimited frequency ranges with the sizes from the same time step for a single battery cell or for several series and / or parallel connections

Battery cells are compared, and that if a difference between the sizes exceeds a threshold value SA, a warning signal WA (battery state change detected) is transmitted to the transmission unit and that if the difference between the sizes exceeds a further threshold value SB, a warning signal WB

(safety-critical battery status) is transferred to the transmission unit.

18. The method as claimed in one of the preceding claims, characterized in that characteristic quantities of the relaxation time spectrum, such as the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, Number of maxima, the distribution of the maxima in the spectrum, the difference between the relaxation times with respect to the maxima of the distribution function of the polarization resistances, the mean value of the maxima, the focus of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the spectrum over one specific frequency range, the integral values over several delimited frequency ranges and the mean value of the integrals over one or more delimited frequency ranges are compared with quantities from at least one previous time step for a single battery cell or for several battery cells connected in series and / or in parallel, and that if a difference between the quantities exceeds a threshold value SA, a warning signal WA (battery state change detected) is transmitted to the transmission unit and that if the difference between the quantities exceeds a further threshold value SB, a warning signal WB (safety-critical battery state) is transmitted to the transmission unit becomes.

19. The method according to any one of the preceding claims, characterized in that that characteristic quantities of the relaxation time spectrum such as the value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima , Number of maxima, the distribution of the maxima in the spectrum, the difference between the relaxation times with respect to the maxima of the distribution function of the polarization resistances, the mean value of the maxima, the focus of the spectrum, the value of the spectrum at a specific frequency, the value of the integral of the spectrum above a specific frequency range, the integral values over several delimited frequency ranges and the mean value of the integrals over one or more delimited frequency ranges with quantities from at least one previously performed measurement for a single battery cell or for several series and / or parallel connected battery cells and that if a difference between the sizes exceeds a threshold value SA a warning signal WA

(Battery state change detected) is transferred to the transmission unit and that if the difference between the sizes exceeds a further threshold value SB Warning signal WB (safety-critical battery status) is transferred to the transmission unit.

20. The method according to any one of the preceding claims, characterized in that statistical variables of the relaxation time spectrum, in particular mean and

Scattering measures for the parameter, value of the global maximum, frequency of the global maximum, the values of the local and / or global maxima, the frequency of the local and / or global maxima, number of maxima, the distribution of the maxima in the spectrum, the difference of the Relaxation times regarding the maxima of the distribution function of the

Polarization resistances, the mean of the maxima, the center of gravity of the spectrum, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a certain frequency range, the integral values over several defined frequency ranges and the mean value of the integrals over one or more defined frequency ranges Sizes from the same time step for a single battery cell or for several battery cells connected in series and / or in parallel are compared, that if a difference between the sizes exceeds a threshold value SA, a

Warning signal WA (battery state change detected) is passed to the transmission unit and that if the difference between the sizes is another

Threshold value SB exceeds, a warning signal WB (safety-critical battery condition) is transferred to the transmission unit.

21. The method according to any one of the preceding claims, characterized in that statistical variables of the relaxation time spectrum, in particular mean and

Polarization resistances, the mean of the maxima, the center of gravity of the spectrum, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a certain frequency range, the integral values over several defined frequency ranges and the mean value of the integrals over one or more defined frequency ranges Variables from at least one previous time step for a single battery cell or for several battery cells connected in series and / or in parallel are compared so that if a difference between the variables exceeds a threshold value SA, a warning signal WA (battery state change detected) is sent to the Transmission unit is transferred and that if the difference between the sizes exceeds a further threshold value SB, a warning signal WB (more critical

Battery status) is transferred to the transmission unit.

22. The method according to any one of the preceding claims, characterized in that statistical variables of the relaxation time spectrum, in particular mean and

Polarization resistances, the mean of the maxima, the center of gravity of the spectrum, the value of the spectrum at a certain frequency, the value of the integral of the spectrum over a certain frequency range, the integral values over several defined frequency ranges and the mean value of the integrals over one or more defined frequency ranges Variables from at least one previously performed measurement for a single battery cell or for several series and / or parallel connections

(safety-critical battery status) is transferred to the transmission unit.

23. The method according to any one of the preceding claims, characterized in that the comparison of a first battery cell with at least one diagnostic size of one or more further battery cells, in particular with spatially immediately adjacent and / or at a distance of at most 50 cm, in particular 10 cm, positioned battery cells takes place within a battery system, in particular in order to detect thermal propagation of a safety-critical state.

24. Excitation unit for carrying out the method according to one of the

The preceding claims in particular comprising a storage capacitor (190), a bidirectional power electronic converter (200) and a filter (210) with a means for charging the storage capacitor before the start of a measurement, the storage capacitor (190) supplying energy by means of a mean-value-free current excitation signal to one or more by means of the bidirectional, power electronic converter

Imprint battery cells so that energy between the storage capacitor and battery can be cyclically shifted, in particular set up to carry out a

Method according to one of the preceding claims.

25. Battery system with an excitation unit according to the preceding claim.

26. Device for performing a method according to one of the preceding claims, comprising a control unit (10), an excitation unit (20), a measuring unit (30), an evaluation unit (40), and in particular one

Control unit (50), a diagnostic unit (60) and / or a transmission unit (70), the control unit (10) in particular for generating a

Excitation signal construction (100) with suitable parameters, in particular

Excitation frequencies, excitation amplitudes, phase positions and one

Overall amplitude scaling (scaling factor) is set up, and in particular the excitation unit (20) is further equipped such that it enables adaptive procurement of the excitation signal construction (100), and in particular is set up to adaptively adapt at least one of the parameters of the excitation signal construction (100).

27. The device according to the preceding claim, characterized in that the measuring unit (30) for simultaneous detection and recording of the

Voltage measurement signal is set up in a previously defined measurement time window from one or more simultaneously excited battery cells.

28. Device according to one of claims 26 or 26, characterized in that the control unit (10) is set up for generating an excitation signal construction of a characteristic multifrequency periodic, continuous signal defined in a control unit in one or at the same time a plurality of battery cells connected in series and / or in parallel .

29. Device according to one of claims 26 to 28, characterized in that it contains a control unit that is set up in such a way that it uses at least one evaluation variable to determine adaptive correction parameters for the amplitudes, frequencies and relative phase positions for the next measurement time window and to transmit them to the control unit (10).

30. Device according to one of claims 26 to 29, with an excitation unit (20) for applying a current excitation signal; with a means for recording an impedance spectrum of the battery cell; with an evaluation unit (40) (means for determining an evaluation variable) based on the measured impedance spectrum), characterized in that the excitation unit (20) (the means for applying a current excitation signal) is designed such that the current excitation signal consists of at least two periodic signals with different ones Frequencies, the current excitation signal can be applied so that it is free of mean values over at least one period of the smallest frequency contained, the measuring unit (30) and the evaluation unit (40) (the means for recording the

Impedance spectrum of the battery cell) are designed so that a first

Response measurement signal of the battery can be determined and evaluated and that at least one of the parameters amplitude, frequency and relative phase position of at least one of the components of the current excitation signal can be changed as a function of the first response measurement signal, so that a measurement error is minimized, that a further response measurement signal is determined and evaluated is that the evaluated value can be used as an evaluation variable and that the determination of the

Battery state of the battery cell based on a comparison of at least one

Diagnostic size can be done with at least one reference value.

31. Battery system with a device according to one of claims 26 to 30.