Method for monitoring an energy storage device in a motor vehicle

Abstract

The invention relates to a method for monitoring an energy store (16, 32, 42) in a motor vehicle, which energy store (16, 32, 42) supplies at least one, in particular safety-relevant consumer (36, 46), preferably for autonomous driving functions, in which at least one characteristic quantity (Up) of the energy store (16, 32, 42) is predicted, the prediction of the characteristic quantity (Up) being made depending on the ratio of the internal resistance (Ri) and the polarization resistance (Rpol) of the energy store (16, 32, 42).

Description

Technical Field

[0001] The present invention relates to a method for monitoring an energy storage device in a motor vehicle based on the field of the independent claims.

Background Art

[0002] DE102019219427A1 relates to a method for monitoring an energy storage device in a motor vehicle, which energy storage device preferably supplies at least one consumption device related to safety, especially for an autonomous driving function, and at this time, depending on the load transition, at least one characteristic quantity of the energy storage device is predicted, thereby determining at least one performance of the energy storage device, and at this time, it is determined whether the energy storage device has been replaced, and when the replacement of the energy storage device is recognized, it is determined whether the replaced energy storage device is an acceptable energy storage device.

Summary of the Invention

Problems to be Solved by the Invention

[0003] The problem underlying the present invention is to further improve the safety and reliability of the on-board network. This problem is solved by the features of the independent claims.

Means for Solving the Problems

[0004] The prediction of the characteristic quantity can be further improved by being performed depending on the ratio of the internal resistance and the polarization resistance of the energy storage device. This is very important especially for applications related to safety during autonomous driving. Furthermore, in addition to easily considering the aging behavior of the energy storage device, this corresponding resistance ratio guarantees a margin for prediction. Since the storage of this ratio is performed depending on a specific measured quantity or state quantity, then, for prediction, the corresponding resistance ratio is based on considering the actual ambient conditions. Thereby, by storing various types of energy storage devices in various ambient conditions, appropriate parameters can also be used during prediction.

[0005] Particularly useful is that measurements are carried out in advance to determine the internal resistance and / or the polarization resistance depending on various ambient conditions or state variables. These measurements can be carried out offline on various types of energy storage devices before their original use during the driving operation, thereby further increasing the accuracy, simulating extreme situations, and reproducing the effects.

[0006] In a useful variant, a load profile is used to determine the internal resistance and / or the polarization resistance, which load profile includes a configurable peak load at which the energy storage device reaches at least a certain voltage. Thereby, worst-case operating cases can already be simulated in advance, and this extreme load profile does not have to be activated later during the driving operation for the prediction of characteristic variables. This increases the operational safety during the driving operation.

[0007] In a useful variant, the polarization resistance is determined depending on the voltage difference that occurs after the application of the load profile and / or depending on the current difference and / or depending on the internal resistance. Thereby, depending on the type of stimulus, preferably with corresponding measured values in the case of various ambient conditions or state variables, the resulting polarization resistance can be determined with high reliability.

[0008] In a useful variant, the predicted characteristic variable is determined by multiplying the internal resistance determined at that time by the ratio of the polarization voltage and / or the pre-stored polarization resistance to the internal resistance and multiplying by a characteristic variable representing the peak load of the load profile. Thereby, in particular, the voltage occurring in the energy storage device is predicted, which voltage depends on the above-mentioned quantities, and the resistance ratio affects these quantities accordingly. In addition, it is particularly useful that the rest voltage and / or the voltage drop at the internal resistance are also taken into account for predicting the characteristic variable. Thereby, the state of charge of the energy storage device and the ambient conditions at that time are also correspondingly taken into account for the prediction, thus further improving the quality of the prediction.

[0009] It is particularly useful for the extreme values, especially the maximum value, to be preserved from the ratio of the polarization resistance to the internal resistance, whereby only the maximum load cases that the energy storage device must surely overcome during operation are accurately based on the prediction. This allows various types of energy storage devices to be easily reproduced without overly increasing the memory requirements.

[0010] In a useful variant, the determined ratio of the polarization resistance to the internal resistance is stored in a correspondence table and / or stored depending on at least one function. Thereby, the memory requirements can be further reduced.

[0011] In a useful variant, the prediction of the characteristic quantity is performed using a quantity that reproduces the influence of the load history of the energy storage device, which is stored in advance depending on at least one state quantity, and this quantity is selected depending on the state quantity. Thereby, the dynamic load or load history of the energy storage device can also be easily considered, thus further improving the quality of the prediction.

[0012] Further useful variants will be apparent from the further dependent claims and the description of the invention.

Brief Description of the Drawings

[0013]

Figure 1

Figure 2

Figure 3A

Figure 3B

Figure 3C

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Mode for Carrying Out the Invention

[0014] The present invention is schematically shown in the drawings based on embodiments, and will be described in detail below with reference to the drawings. In an exemplary embodiment, a battery or a storage battery is exemplarily described as an energy storage device that can be used. However, instead, other energy storage devices suitable for this problem, such as those based on inductance or capacitance, fuel cells, capacitors, or the like, can be used in the same way.

[0015] FIG. 1 shows a possible topology of an energy supply system consisting of a base on-board network 22 that supplies at least one base consumer device 24, illustratively shown. As an alternative, an energy storage device or battery and / or a starter with an attribution (battery) sensor and / or a plurality of comfort consumer devices not related to safety may be provided within the base on-board network 22, and these may be secured or controlled by electrical load distribution. The base on-board network 22 has a voltage level lower than that of the high-voltage on-board network 10 and may be, for example, a 14V on-board network. A DC voltage converter 20 is arranged between the base on-board network 22 and the high-voltage on-board network 10. The high-voltage on-board network 10 illustratively includes a high-voltage energy storage device 16, for example a high-voltage battery, which may incorporate a battery management system in some cases, and a load 18 not related to safety, shown as a specific example, that is, a comfort consumer device such as an air conditioner supplied with an elevated voltage level, and an electric machine 12. The energy storage device 16 may be connected via switching means 14 to supply the high-voltage on-board network 10. High voltage, in this context, means a voltage level higher than that of the base on-board network 22. That is, it may be, for example, a 48-volt on-board network. Alternatively, in the case of a vehicle equipped with an electric powertrain, it may be a higher voltage level. Alternatively, the high-voltage on-board network 10 may be completely eliminated, and in this case, components such as a starter, a generator, and an energy storage device are assigned to the base on-board network 22.

[0016] The base on-board network 22 is connected to, for example, two safety-related channels 30, 40. The first safety-related channel 30 is connected to the base on-board network 22 via a separation element 28. A further safety-related channel 40 is connected to the base on-board network 22 via a further separation element 26. The first safety-related channel 30 can be supplied with energy by an energy storage device 32. Characteristic quantities of the energy storage device 32 are captured by a sensor 34. The sensor 34 is preferably arranged adjacent to the energy storage device 32. The first safety-related channel 30 supplies a safety-related consumer device 36. This safety-related consumer device 36 is merely shown as an example. If necessary, other safety-related consumer devices 36 are also supplied by the safety-related channel 30.

[0017] The further safety-related channel 40 can also be supplied with energy by a further energy storage device 42. Characteristic quantities of the further energy storage device 42 are captured by a further sensor 44. The further sensor 44 is arranged adjacent to the further energy storage device 42. This further safety-related channel 40 supplies at least one further safety-related consumer device 46. If necessary, other safety-related consumer devices 46 can also be supplied within this further safety-related channel 40.

[0018] The topology shown in FIG. 1 has been selected merely by way of example as one of many exemplary embodiments. There are very diverse possibilities for attaching the safety-related channels 30, 40. By way of example, the further safety-related channel 40 could be connected to the safety-related channel 30 or connected to channel 10 via a further DC voltage converter.

[0019] The separation elements 26, 28 serve to ensure the safety of the channels 30, 40 related to their respective safety, whereby defects occurring in particular within the base on-board network 22 and / or within one of the channels 30, 40 related to safety cannot affect the other channel 30, 40 related to safety. This can be corresponding switching means or even a DC voltage converter that enables the separation or connection of sub-networks. Instead, the separation elements 26, 28 could be completely eliminated, whereby the channels 30, 40 are directly connected to the DC voltage converter 20.

[0020] The redundant, in particular functionally redundant, safety-related consumer devices 36, 46 that can be supplied by both channels 30, 40 related to safety are the consumer devices required to transfer the vehicle from an autonomous driving operation (without the need for driver intervention) to a safe state, for example in the case of a critical defect. A safe state can be, immediately or eventually, a stop of the vehicle, at the roadside or at the nearest rest area, for example.

[0021] Nevertheless, the functional capabilities of the energy storage devices 16, 32, 42 for supplying one or more safety-related consumer devices 36 play an important role even in the case of possible defects. With the introduction of electrical steering and braking as well as the evolving automation of vehicles, it is becoming increasingly important to ensure a reliable electrical supply to these safety-related components or consumer devices 36, 46. Since the energy storage devices 16, 32, 42 play a decisive role in this regard, functions that will surely determine the performance of the energy storage devices 16, 32, 42 must now be developed based on particularly high requirements, such as those described in ISO 26262. This has far-reaching consequences for the development of functions and algorithms as well as for the hardware to which these functions are applied. The method described below enables a reliable prediction of characteristic variables, such as the voltage Up of the energy storage devices 16, 32, 42, based on ISO 26262. For a safe on-board network 30, 40 based on safety criteria, the prediction of the characteristic variable Up of the energy storage devices 16, 32, 42 is an essential component.

[0022] When predicting performance in the context of reliably supplying safety-related consumer devices 36, 46, it is important to ensure that the energy storage devices 16, 32, 42 can supply at least one or more safety-related consumer devices 36, 46 until the vehicle is in a safe state (the vehicle is safely at the roadside, in a parking space, etc.), which is evident, for example, from the overlap of the steering process and the braking process. The load estimation of the safety-related consumer devices 36, 46 is represented by the (current) profile of the so-called State-of-Function (SOF). Such a load profile 50 is shown as a specific example in FIG. 2 in the form of the essential current I required therefor by the energy storage devices 16, 32, 42. In this regard, each load profile 50 can define the maximum demand on the energy storage devices 16, 32, 42, and in this case, when implementing the load profile 50, the characteristic variable Up must not drop below a certain limit value Ulimit.

[0023] The particularity of this approach in the context of security ultimately lies in the fact that it cannot be assumed that it is known which energy storage devices 16, 32, 42 will be used. If the replacement of the energy storage devices 16, 32, 42 is not monitored, for example, there may be a situation where energy storage devices 16, 32, 42 that have not been previously measured and whose characteristics are unknown are used. Therefore, a further important part of this solution is to generate parameter sets 78, 80 that cover a large number of energy storage devices 16, 32, 42 without violating the safety objectives.

[0024] The measurement concept for determining the parameter sets 78, 80 is designed generically, so that a very large number of different SOF profiles (customer requirements) can be created by one corresponding parameterization. The envelope 52 around the load profile 50 provides a generalized approach that enables a flexible solution for the customer (such as the current profiles in FIGS. 3A, 3B, 3C).

[0025] Since the solution space can become very large, it is important to save a large number of parameter values 78, 80. The computing resources and memory resources in the vehicle are generally very limited, especially in the case of components such as sensors 34, 44, especially battery sensors that have to fit, for example, in the battery slot, so this is no longer possible for memory reasons. Therefore, regression is performed so that the required amount of memory is kept as small as possible. A lookup table or correspondence table 72 can be used as a solution. An alternative approach is to further significantly reduce the amount of memory by a function such as a polynomial function 76 instead of the lookup table or correspondence table 72.

[0026] In energy storage devices 16, 32, 42, an important quantity for predicting whether the energy storage device can provide the required output is, in the case of a rechargeable battery as the possible energy storage devices 16, 32, 42, the polarization resistance Rpol in addition to the internal resistance Ri. The determination of the polarization resistance Rpol is a complex quantity based on the multi-factor dependence of the polarization resistance (time t, current level I, temperature T, state of charge SOC of the energy storage device, type and progression of aging, structural form and structural type of the energy storage device, preloading), and is difficult to determine.

[0027] This means, with respect to a function or algorithm, that in a safety-related context, the algorithm must recognize the multi-factor relationships for all possible operating scenarios and must weight the multi-factor relationships appropriately. In this regard, a reliable prediction of the important characteristic quantity Up of the energy storage devices 16, 32, 42 must always be made. For example, the aging of the energy storage devices 16, 32, 42 must be taken into account.

[0028] In a safety context, what can be very well determined is the ohmic internal resistance Ri of the energy storage device 16, 32, 42 or the battery. This ohmic internal resistance Ri changes with the aging of the rechargeable battery and is thus, if not the only one, an indicator regarding aging. Nevertheless, the change in the value of the internal resistance Ri can be used for a reliable prediction of the performance of the energy storage device 16, 32, 42 during the service life of the energy storage device 16, 32, 42, whether the energy storage device 16, 32, 42 still exists or no longer exists.

[0029] This solution is based on the fact that the ratio of the polarization resistance Rpol to the internal resistance Ri, which can be determined very well at the beginning of the life of the energy storage devices 16, 32, 42, changes characteristically during the course of aging. In particular, it can be assumed that the increase in the internal resistance Ri due to aging is always the same as or even greater than the increase in the polarization resistance Rpol due to aging (Figure 7). By determining the ratio of these two quantities Ri, Rpol and maintaining this ratio even when the energy storage devices 16, 32, 42 age, a safety margin is obtained, which is suitable for predictions from the perspective of safety, for example, based on the standard ISO 26262.

[0030] Figure 2 shows, by way of example, the time course of the load profile 50. This load profile 50 reproduces, by way of example, the possible typical load transitions of the safety-related consumer devices 36, 46, which load transitions are called up, for example, in the case of defects, in particular in typical situations such as the implementation of a stop on the hard shoulder, or similar, in any case as a minimum requirement for further driving operations to be carried out. The sometimes strongly fluctuating time course of the load profile 50 is approximated by the envelope 52. This approximation is carried out in view of the guarantee of functionality, such that in case of doubt, a higher load in the energy storage devices 16, 32, 42 is called up for the prediction based on this when using the envelope 52. In this exemplary embodiment, the current I to be provided by the energy storage devices 16, 32, 42 is used as the load profile 50.

[0031] To predict the output or characteristic quantity Up (so-called State-of-Function SOF) of the energy storage devices 16, 32, 42 or the battery, a one-stage or two-stage load profile 50 with a defined length of time, for example a current profile, is defined, and this load profile 50 is used to form an envelope line 52 around the estimated load profile 50 in the case of safety. The load profile 50 or the attributable envelope line 52 can be varied in terms of values, in particular current values, and duration according to the customer's wishes. This solution can be used flexibly for various current profiles (Figs. 3A to 3C). The load profile 50 can be flexibly adapted according to the customer's wishes. However, there are certain maximum limits that depend on the capacity of the energy storage devices 16, 32, 42, in particular the battery capacity. That is, the maximum limit can be defined, for example, within a certain time span (tpeak, for example, up to 15 seconds) for a certain peak load 53 (Ipeak, for example, 5 times the battery capacity C per hour, for example 300 A), and within 60 seconds for a base load 51 (for example, up to 2 times the battery capacity C per hour, for example 120 A). These exemplary values can be read from Fig. 3.

[0032] This defined and customer-adjusted load profile 50 or envelope of attribution 52 is used, as shown in FIG. 4, for output prediction, i.e., for the prediction of characteristic quantities such as the voltage Up (on-board network voltage, voltage at the energy storages 16, 32, 42). In this case, the load profile 52 includes, if necessary, the base load 51 and the peak load 53. The peak load 53 is defined by the maximum load, for example the maximum current Ipeak, which is generally applied for a shorter duration tpeak than the base load 51. The base load 51 is generally characterized by a longer load progression at a lower load. If necessary, further quantities such as the temperature T or the current I at the energy storages 16, 32, 42 are measured, for example by the respective sensors 34, 44. As will be explained in more detail in FIGS. 5 and 6 later, the defined load profile 50 or envelope of attribution 52 is used as a load basis for the prediction of the characteristic quantity Up that will occur, for example the voltage prediction.

[0033] In an exemplary embodiment based on FIG. 5, at block 62, measurements of various energy storage devices 32.1, 32.2... 32.n are performed. For this, the general measurement specification 60 provided to the measurement block 62 is used. These measurements for the various energy storage devices are performed offline, i.e., before the start of the method, for example, in sensors 34, 44, preferably outside the vehicle. During this measurement, the customer-specific load profile 50 as exemplified in connection with FIGS. 2 and 3 is used as a basis. Substantially, the values of current and voltage occurring at the beginning and / or end of the edge change in the load profile 50 are captured. Regarding lead-acid batteries as possible energy storage devices 16, 32, 42, initially, with no load in the energy storage devices 16, 32, 42, a certain voltage U0, also called the OCV open circuit voltage or rest voltage, occurs. When a current profile or current edge is applied, the voltage response first transitions to the height of Ri*I in an immediate, time-independent voltage drop (voltage drop based on the internal resistance Ri). Subsequently (after the almost immediate voltage drop to the height of Ri*I), the further voltage drop Upol follows a strong dependence on the time t of the load profile 50 until a further voltage drop Upol to the height of the polarization voltage Upol occurs based on a certain time constant. The voltage value (Up = U0 - I*Ri - Upol) that would occur in the case of this load profile (I) must be greater than the minimum allowable voltage Ulimit, thereby ensuring more reliable operation of the energy storage devices 16, 32, 42.

[0034] The measurements based on block 62 are repeated for each energy storage device 32.1, 32.2, 32.3, 32.4... 32.n. In this case, the corresponding load profile 50 or the surrounding line 52 of attribution as described in connection with FIGS. 2 to 4 is used as a basis.

[0035] This allows this concept to offer the possibility that model 86 is adapted (Fig. 6) either to a customer-specific solution based on the corresponding parameter set of a particular underlying battery (see block 75 in Fig. 5) or generalized for any number of selected batteries (see block 74 in Fig. 5). In this case, the capture of battery measurement data according to general measurement specifications 60 is used as a basis. Using this (battery) measurement, the individual influencing factors of the voltage behavior can be analyzed for each battery. The parameterization encompasses, among other things, the dependence of the polarization voltage Upol on the internal resistance Ri of the energy storage devices 16, 32, 42 or the battery. This relationship is approximated and used in the form of an analytical function. The parameterization of model 86 is carried out for a specific load profile 50 or envelope 52.

[0036] From these measured values U, I or the known load profile 50, a regression analysis is carried out for each battery (block 64). Thus, the measurements for determining the internal resistance Ri in block 66 are repeated for various ambient conditions and various energy storage devices 16, 32, 42. The ambient conditions can be varied. This can be, for example, the open-circuit voltage U0 and / or the temperature T of the energy storage devices 16, 32, 42. However, instead, other influencing quantities or quantities related thereto, such as the state of charge SOC or the like, can also be used. Thus, in this exemplary embodiment, for each measured energy storage device type, a two-dimensional Ri field depending on the open-circuit voltage U0 and the temperature T is created (Ri(U0,T)). As a stimulus for the measurement, as described, a load profile 50 or an envelope 52 of attribution that can reproduce a worst-case scenario and can be individually set by the customer is used.

[0037] In block 68, the polarization resistance Rpol is determined, which also depends on specific ambient conditions. In this exemplary embodiment, the ambient conditions such as current I and temperature T are used, which are precisely varied for each energy storage type, and the polarization resistance Rpol is determined under the varied conditions. The rest voltage U0 may also be used as a further ambient condition. In principle, the determination of the polarization resistance Rpol is performed using the following equation.

[0038] [Number]

[0039] where U0 is the rest voltage, U1 is the voltage value occurring at the end of the load profile 50, I0 is the current used at the beginning of the load profile 50, and I1 is the current used at the end of the load profile 50, and Ri is the internal resistance.

[0040] Thereby, the polarization resistance Rpol for each energy storage type at the varied load profile 52 or current profile and the varied temperature profile T is captured. Thus, in this exemplary embodiment, for each measured energy storage type, a three-dimensional Rpol field (Rpol(I,T,U0)) that depends on the current I of the respective maximum current Ipeak is created for the respective load profile 52, temperature T, and rest voltage U0.

[0041] The output powers of blocks 66 and 68 reach block 72 (via blocks 74 and / or 75 as briefly mentioned above), where the ratio (Rpol / Ri) of the polarization resistance Rpol to the internal resistance Ri is formed, in particular in the form of a correspondence table. This ratio Rpol / Ri is formed depending on the battery temperature T and on the open-circuit voltage U0 and is stored in the form of a correspondence table. For each battery type, the corresponding ratio Rpol / Ri of the individual values Rpol(U0,T) or Ri(U0,T) for the same open-circuit voltage U0 or the same temperature T is determined. This function can be adapted to one particular battery solution 32.1 or 32.2 or 32.3 etc. (block 67) or to the entire group of batteries or energy storage devices 16, 32, 42 (block 69). In block 69, in the sense of a safe design, it is preferably the maximum value of Rpol / Ri for each battery group 16, 32, 42 that is selected for each operating point. This (maximum) ratio Rpol / Ri is stored, for example, depending on any number of open-circuit voltage values U0 (e.g., between 11.5 and 13 V) and temperature values T (e.g., between -20 °C and 70 °C) and can be called from the correspondence table 72 as a value pair.

[0042] Instead, the worst-case value pairs can be reproduced by an analytical function, for example a polynomial function or other suitable function. Thus, the memory consumption for storing the model behavior can be further reduced. This is done in block 76.

[0043] Thereby, the value pairs (Rpol / Ri) as a function f=(T,U0) depending on the temperature T and / or the open-circuit voltage U0 are provided as parameter 78, either in the form of the correspondence table 72 or in the form of a generalization by a function or polynomial based on block 76. This concludes the offline parameterization. Parameter 78 is provided for further use during operation of the vehicle as will be explained below with reference to FIG. 6.

[0044] In FIG. 6, the evaluation of the respective states of the energy storage devices 16, 32, 42 during the operation of the vehicle is explained. This evaluation can be realized, for example, in the sensors 34, 44. Instead, this could also be done in further control devices within the vehicle or outside the vehicle, for example, in the cloud.

[0045] Based on FIG. 6, various measurement data such as the voltage U, current I, temperature T, as captured by the sensors 34, 44, lead to a state recognition 82 for the energy storage devices 16, 32, 42. The state recognition 82 can be realized in the sensors 34, 44, or instead in further control devices within the vehicle or outside the vehicle, for example, in the cloud.

[0046] In the state recognition 82, characteristic quantities of the energy storage devices 16, 32, 42, that is, state quantities such as the internal resistance Ri, the temperature T of the energy storage devices 16, 32, 42, the rest voltage U0, and optionally further quantities, are determined using the measured quantities U, I, T and, if necessary, a model or the like on which it is based. The state recognition 82 can determine the temperature T of the energy storage devices 16, 32, 42 as the state quantity T based on a specific algorithm, for example, based on the measured ASIC temperature T (measured quantity). Certain state quantities, such as the internal resistance Ri, the temperature T, the rest voltage U0, lead to a static model 86. Parameters 78 and / or load profiles 50, 52, especially the current profile, as determined in FIG. 5, are sent to this static model 86. This static model 86 determines the (static) polarization voltage Upol via the ratio of the polarization resistance Rpol to the internal resistance Ri (Rpol / Ri). To determine the internal resistance Ri by the state recognition 82, a small stimulation of the on-board network is required by accurately connecting a load if necessary. The main feature of this approach is that the logic for prediction does not require a load profile similar to the load profile 52 with the maximum current height Ipeak or the maximum duration tpeak.

[0047] The polarization voltage Upol in the energy storage devices 16, 32, 42 can be determined by the following equation. For this purpose, the corresponding ratio Rpol / Ri, which depends on the temperature T determined by the sensors 34, 44 or the state recognition 82 or the rest voltage U0 determined by the state recognition 82, is determined from the look-up table 72 or the polynomial 76 (in the case of maximum load (Ipeak, tpeak)). As the current I, the maximum current Ipeak from the load profiles 50, 52 respectively based on in FIG. 4 is used. The internal resistance Ri and the rest voltage U0 are provided by the state recognition 82. Thereby, the polarization voltage Upol can be determined using the following second equation.

[0048] Rpol = f(t peak , I peak , U0, T) Ri = f(U0, T)

[0049]

Equation

[0050] The voltage drop U across the internal resistance Ri is calculated by the following formula. U Ri (T, U0) = Ri(U0, T)·I peak Wherein, Ri(U0, T) is determined by the state recognition 82 based on the measured current values and voltage values I, U, and filtered as necessary.

[0051] In the subsequent block 90, the voltage Up occurring in the energy storage devices 16, 32, 42 is predicted. This prediction in block 90 is performed based on the following equation depending on the (steady-state) polarization voltage Upol, the rest voltage U0, and the voltage drop U Ri across the internal resistance.

[0052] U p = U0 - U Ri - U pol When the predicted voltage Up falls below the voltage limit value Ulimit, for example 9V, a warning or countermeasure is initiated (Up < Ulimit). In this case, the energy storage devices 16, 32, 42 are not considered to be performing anymore.

[0053] Thus, model 86 uses the determined parameters 78 of the respective energy storage devices 16, 32, 42 together with the current profile or load profile 50, 52 as input data (Figure 6).

[0054] The function values are always adapted in the sensors 34, 44 by the changed operating conditions, such as the temperature T of the energy storage devices 16, 32, 42 or the changed state of charge SOC or the rest voltage U0. The adaptation of the voltage prediction Up for aging is inherently ensured by the ratio of the polarization resistance Rpol to the internal resistance Ri as detailed below, and thus adaptation for aging is not necessary.

[0055] The aging mechanisms that very often occur in lead-acid batteries as possible energy storage devices 16, 32, 42 are corrosion (1) and loss of active mass (2). (1): Corrosion manifests itself, among other things, by an increase in the internal resistance Ri. Since the ratio (Rpol / Ri) of the polarization resistance of the non-aged energy storage devices 16, 32, 42 or battery to the internal resistance is selected as a parameter set, it is inferred that due to corrosion as the main aging mechanism, the ratio Rpol / Ri decreases due to corrosion, or at least does not increase.

[0056] (2): The same applies to the loss of active mass (LAM). The loss of active mass reduces the reactive active battery surface area, whereby both resistance types Rpol and Ri increase, but the ratio Rpol / Ri does not increase. An increase in Rpol / Ri can only occur for batteries that are very strongly LAM-aged, but in this state, the battery is already so aged that it has to be removed anyway.

[0057] Finally, the polarization resistance Rpol is determined depending on the ratio Rpol / Ri with respect to the beginning of life (BOL) of the energy storage devices 16, 32, 42 of the polarization resistance, by the aged internal resistance Ri_g applied at that time: (Rpol / Ri) BOL *Ri_g.

[0058]

Number

[0059] In FIG. 7, the relationship described depending on time t, i.e., aging, is shown explicitly. That is, the internal resistance Ri at the end of life (EOL) based on the corrosion of the energy storage devices 16, 32, 42 increases disproportionately large compared to the polarization resistance Rpol. Considering the ratio Rpol / Ri (Kor) based on corrosion, this ratio decreases linearly. Different from corrosion, the polarization resistance Rpol based on the loss of active material (LAM) increases relatively strongly towards the end of life EOL. Also here, the ratio Rpol / Ri (LAM) based on the loss of active material LAM does not increase initially. This ratio finally reaches again the initial level at the beginning of life of the energy storage devices 16, 32, 42 when LAM appears very strongly. After this point, it can be inferred that the end of life (EOL) of the energy storage devices 16, 32, 42 has been reached.

[0060] The exemplary embodiment based on FIG. 9 supplements the procedure for creating the parameter 80 for the dynamic model 88 with respect to the exemplary embodiment of FIG. 5. The dynamic model 88 will be described in detail later in connection with FIG. 10.

[0061] The essential feature of this concept is the consideration of the load history of the energy storage devices 16, 32, 42, especially the battery, for the voltage prediction Up. In the driving operation, there is a dynamic load situation, which also means various loads on the energy storage devices 16, 32, 42 depending on the driving pattern. That is, in the dynamic driving situation under a strong load on the energy storage devices 16, 32, 42 (discharge of the energy storage devices 16, 32, 42), in order to make the voltage prediction Up reliable and trustworthy, not only the static operating point (in the so-called equilibrium state of the energy storage devices 16, 32, 42) needs to be considered, but especially the operating history should be taken into account. The concept introduced here provides the possibility of extending an arbitrarily derived voltage prediction Up (constant estimated values or arbitrary load profiles 50, 52 as exemplified in connection with FIG. 5) for dynamic load cases. Thereby, the voltage prediction Up can be carried out significantly more accurately. One feature of this concept is the temporal weighting of the occurring load, that is, when an arbitrary load occurs on the energy storage devices 16, 32, 42, the predicted value Up (as exemplified in block 90 based on FIG. 6 for the static operating case) is adapted. However, for the static part of the prediction characteristic Up, another prediction or estimation method that does not utilize FIG. 6 is also possible.

[0062] In this case, for the basis, the capture of measurement data of the energy storage devices 16, 32, 42 using a general measurement specification 60 is used. Using these measurements, the influencing factors of the voltage behavior can be analyzed separately for each of the energy storage devices 16, 32, 42. This parameterization also includes the dynamic behavior of the energy storage devices 16, 32, 42. The dynamic resistance behavior of the energy storage devices 16, 32, 42, especially the battery, is approximated in the form of an analytical function and used. When the predicted value is derived via the load profiles 50, 52 as described in connection with FIGS. 2 and 3, the previous load of the battery as an additional input quantity can make the dynamic algorithm even more accurate.

[0063] The height I of the current and the load time t of the load profile 50 are used as the input quantity. Charge accumulation is carried out and reset through a time-dependent decay behavior during the rest period. The time-dependent decay function can be implemented by an e-function or a pT1 element. The analytical function describes the relationship between the charge history Q and the additional overvoltage U. The occurrence of the overvoltage U and the decay behavior are both validated based on measurement data (see the general measurement plan). As a result, the voltage prediction Up can proceed more reliably even in a highly dynamic process. That is, this algorithm weights the predicted value Up depending on the preload of the energy storage devices 16, 32, 42, which is represented by the amount Udyn,τ that reproduces the influence of the load history of the energy storage devices 16, 32, 42 on the characteristic quantity Up.

[0064] The polarization resistance Rpol of the energy storage devices 16, 32, 42 can be very diverse depending on the operating conditions of the energy storage devices 16, 32, 42. Basically, part of the polarization resistance Rpol is always the charge transfer resistance (overvoltage) that depends on the temperature T and the state of charge SOC of the energy storage devices 16, 32, 42. An additional component of the polarization resistance Rpol can be the mass transport limit within the energy storage devices 16, 32, 42, such as a diffusion process. The diffusion effect is substantially affected by the state of the energy storage devices 16, 32, 42. Therefore, in the function of the relationship between the load history and the overvoltage, it is expressed as a function of the temperature T and the state of charge SOC of the energy storage devices 16, 32, 42.

[0065] The voltage prediction Up is extended for a dynamic load situation (Figure 11). In an exemplary embodiment based on Figures 5 and 6, an algorithm is illustrated that can predict the polarization voltage Upol of the energy storage devices 16, 32, 42 depending on the operating point. The polarization voltage Upol is measured under defined ambient conditions (the operating point from rest where the energy storage devices 16, 32, 42 are in a relaxed state). This is important for recognizing the basic dependencies considering the operating conditions (the height Ipeak of the pulse current, the battery temperature T, the duration tpeak of the current pulse Ipeak, the state of charge SOC of the energy storage devices 16, 32, 42, or the rest voltage U0) (Figure 10).

[0066] For various types of energy storage devices 32.1, 32.2, 32.3, 32.4,... 32.n, the voltage transition U of attribution is determined for various load profiles 52. For example, for a load profile having a base load 51 and a peak load 53 once (Figure 3A) as shown in Figure 8, and then for a further load profile having only the peak load 53 without a base load (Figure 3C), the voltage transition U of attribution is determined. What is essential is that there is a pause period, i.e., a relaxation period, between both loads. The rise of the voltage U during the pause period occurs with a relaxation time constant, i.e., a decay behavior τ, which is temperature-dependent τ(T). This method can be applied to several battery types or battery groups. This pause period can be selected to be at least as long as the energy storage devices 16, 32, 42 recover again during it, i.e., a rest voltage U0 occurs again. The first load profile 52 (similarly to the further load profile 52) is applied when the energy storage devices 16, 32, 42 are in a defined state, especially when the rest voltage U0 is applied. In this way, the quantities Udyn, τ are determined, which are stored depending on at least one state quantity U0, T, SOC, Q, Ri and reproduce the influence of the load history of the energy storage devices 16, 32, 42 on the characteristic quantity Up.

[0067] At that time, the peak load 53, as used exemplarily in the load profile 52 or the further load profile 52, is defined by the maximum load applied for a duration tpeak, for example, the maximum current Ipeak. The base load 51 is generally characterized by a longer load transition with a lower maximum load. The application of the base load 51 causes the voltage U to drop, and the application of the peak load 53 causes it to drop further. The quantity of interest is the total voltage drop U DCP that occurs then. After the application of the peak load 53 to the energy storage devices 16, 32, 42 ends, the voltage U rises again after a certain pause time. Thereafter, the energy storage devices 16, 32, 42 are applied with only the peak load 53 without the prior application of the base load 51. At this time, the voltage U is at a lower value U SCPIt decreases. The actual voltage behavior Udyn, that is, the value of the dynamically measured voltage Udyn, is calculated by the following equation.

[0068] Udyn(Q,U0,T)=U DCP -U SCP The above equation can be approximated by terms that depend exponentially on temperature and linearly on charge. The dynamic voltage Udyn depends on the charge Q (charge Q as the integral of the current transition I including the temperature-dependent decay behavior or the load 52), the state of charge SOC or the rest voltage U0, and the temperature T of the energy storage devices 16, 32, 42. The dynamic voltage Udyn can represent a measure of the load history of the energy storage devices 16, 32, 42 as the quantity Udyn, τ that reproduces the influence of the load history of the energy storage devices 16, 32, 42 on the characteristic quantity Up.

[0069] Measurements of the voltage transition are repeated while changing ambient parameters such as the temperature T, the preload Q (charge Q as the integral of the current transition I), and the state of charge SOC or the rest voltage U0. This is done in block 62 based on FIG. 9.

[0070] In the subsequent block 64, regression is performed for each energy storage device 16, 32, 42 based on the measured values for various state quantities Q, SOC, T. Coefficients for the following typical equation can be determined for the dynamic voltage Udyn that depends on the state quantities Q, U0 or SOC, T determined above.

[0071] Udyn(Q,U0,T)=a(U0)*exp(b(U0) / T)*Q+c(U0) In the formula, a, b, c are constant parameters to be determined depending on the measured values. In this case, a, b, c are quantities that depend on the rest state U0 (or the state of charge SOC).

[0072] The relaxation behavior shows how the voltage U returns again when it is applied by the base load 51 after, but before, the end of the full load 53. This can be modeled, for example, by a new state quantity Qbat as shown in FIG. 11. This Qbat reproduces the current integral Q during the discharge period, but reproduces the charge value Q of the time-dependent decay behavior during the rest period. The time-dependent decay behavior τ is temperature-dependent τ(T).

[0073] Qbat = f(I, Q, τ(T)) In the formula, Qbat represents the state quantity of the charge Q regarding the load effectively applied to the battery.

[0074] The analysis of the relaxation behavior can be performed from battery measurements as shown in FIG. 8, for example. The relaxation time τ can be stored depending on the temperature T in a correspondence table (τ(T)), for example. The lower the temperature T, the longer the relaxation time τ. This can be done for one specific (block 75) or each (block 74) energy storage device 16, 32, 42 measured offline.

[0075] Instead, for each type of the energy storage devices 16, 32, 42, a combination of determined measured values (Udyn(Q, SOC, T)) can be stored without regression for the determination of the subsequent online predicted voltage Up (based on FIG. 11 described below).

[0076] Instead of the generalization 74, the parameters 80 can be determined (block 75) within the framework of the solutions specific to each of the energy storage devices 16, 32, 42. It is preferable that these described steps also proceed offline here, that is, be carried out before the start of the operation of this method.

[0077] After the generalization of the parameters performed in block 74, the parameter 80 is determined for the dynamic model 88 based on FIG. 11. In this dynamic model, the determination of the state quantity Qbat and the voltage determination 94 (voltage determination Upol,dyn (predicted dynamic polarization voltage)) are distinguished. The state quantity Qbat uses the charge curve Q (current integration) at that time, the battery temperature T, and the current value I at that time as input data. In block 92 (determination of Qbat), the charge Q in the state where no current is flowing is reset by a time delay element. This Qbat is used in the subsequent prediction Upol,dyn by the above equation.

[0078] The form based on FIGS. 10 and 11 is implemented during the operation of the vehicle. This is done using a dynamic model 88 that reproduces the load history of the energy storage devices 16, 32, 42 by making use of the parameter 80 determined in FIG. 9, among other things. Here too, the sensors 34, 44 determine the corresponding measurement data, such as the voltage U, the current I, and the temperature T of the energy storage devices 16, 32, 42, and provide a state recognition 82 of the energy storage devices 16, 32, 42. The state recognition 82 determines the state quantities of the energy storage devices 16, 32, 42 determined based on the corresponding measurement data, and these state quantities, such as the temperature T of the energy storage devices 16, 32, 42 and / or the charge Q or Qbat of the energy storage devices 16, 32, 42 and / or the internal resistance Ri of the energy storage devices 16, 32, 42 and / or the state of charge SOC and / or the rest voltage U0 are also provided to the dynamic model 88. Furthermore, the dynamic model 88 obtains the parameter 80 determined in FIG. 9. If necessary, the load profiles 50, 52 are provided to the dynamic model 88. The dynamic model 88 predicts the resulting dynamic voltage Upol,dyn as a measure of the load history.

[0079] Thus, illustratively in FIG. 10, the total polarization voltage Upol is determined from the sum of the static polarization voltage Upol,stat (or Up,stat based on FIG. 10) and the dynamic polarization voltage Upol,dyn (or Up,dyn based on FIG. 10).

[0080] Upol = Upol,stat + Upol,dyn Upol,dyn = a(U0) * exp(b(U0) / T) * Qbat + c(U0) The predicted characteristic quantity Up results from the following equation.

[0081] Up = U0 - Upol - U Ri In the formula, U0 is the rest voltage, Upol is the polarization voltage, and U Ri represents the voltage drop across the internal resistance.

[0082] When the predicted voltage Up falls below a critical limit value Ulimit (Up < Ulimit), this indicates that the energy storage devices 16, 32, 42 are no longer healthy, and corresponding countermeasures or warning instructions are initiated.

[0083] It is also mentioned here that, if necessary, constant quantities, such as Upol,stat, can be determined in a way different from that described in connection with FIGS. 5 and 6, or the dynamic quantity Upol,dyn can already, by itself, enable an indication of the quality of the energy storage devices 16, 32, 42.

[0084] If the actual functional capabilities of the energy storage devices 16, 32, 42 are not achieved, countermeasures are initiated. That is, for example, a warning notification is issued and / or safety-related functions are blocked. The warning notification can be displayed for the vehicle driver within a display or other display means. Instead, the corresponding warning notification may be displayed via an appropriate communication channel to, for example, a repair shop, the owner of the vehicle fleet, etc. It may be initiated to manually or automatically transfer the vehicle to a safe state, such as a stop at the roadside, driving to the nearest parking space or a similar location (the so-called safe stop of the vehicle).

[0085] The described method is particularly suitable for monitoring energy storage devices 16, 32, 42 for use in safety-related applications, for example, in particular, for supplying consumer devices related to safety in a motor vehicle during autonomous driving. However, the use is not limited to this.

Claims

1. A method for monitoring an energy storage device (16, 32, 42) in a motor vehicle, wherein the energy storage device (16, 32, 42) supplies at least one safety-related consumer device (36, 46) for an autonomous driving function, and at this time, in a method in which at least one characteristic quantity (Up) of the energy storage device (16, 32, 42) is predicted, the prediction of the characteristic quantity (Up) is performed depending on the ratio of the internal resistance (Ri) and the polarization resistance (Rp) of the energy storage device (16, 32, 42), and at least one of the energy storage device (16, 32, 42) at that time. Depending on the state quantity (T, U0, SOC) and / or at least one of the measured quantities (U, I, T) of the energy storage device (16, 32, 42) at that time, the ratio of the polarization resistance (Rp) to the internal resistance (Ri) is selected, and the method used for the prediction of the characteristic quantity (Up).

2. Beforehand, measurements for determining the internal resistance (Ri) and the polarization resistance (Rp) are performed depending on ambient conditions or state quantities (T, U0, SOC) including at least the temperature (T), state of charge (SOC), or rest voltage (U0) of the energy storage device (16, 32, 42), and the ratio of the internal resistance (Ri) and the polarization resistance (Rp) is determined. The method according to claim 1, characterized in that it is carried out.

3. A load profile (50) is used to determine the internal resistance (Ri) and / or the polarization resistance (Rp), the load profile (50) includes a configurable peak load (53), and at the peak load (53), the energy storage device (16, 32, 42) reaches at least a certain voltage (Ulimit). The method according to claim 2, characterized in that it does.

4. The polarization resistance (Rp) is determined depending on the voltage difference (U1 - U0) that occurs after the application of the load profile (50), and / or depending on the current difference (I1 - I0), and / or depending on the internal resistance (Ri). The method according to claim 3, characterized in that it is.

5. The method according to claim 3, characterized in that the determined internal resistance (Ri) is multiplied by the predicted characteristic quantity (Up) which depends on the ratio to the polarization voltage (Upol) and / or the internal resistance (Ri) of the polarization resistance (Rpold) stored in advance, and is multiplied by the peak current (Ipeak) which is a characteristic quantity representing the peak load (53) of the load profile (50).

6. In order to predict the characteristic quantity (Up), the polarization voltage (Upol) and / or the voltage drop (U Ｒｉ ) at the internal resistance (Ri) are subtracted from the rest voltage (U0). The method according to claim 5, characterized in that.

7. The method according to claim 1, characterized in that the voltage (U) in the energy storage device (16, 32, 42) and / or the current (I) applying a load to the energy storage device (16, 32, 42) and / or the temperature (T) of the energy storage device (16, 32, 42) are captured by sensors (34, 44).

8. The state recognition (82) of the energy storage device (16, 32, 42) realized in the sensors (34, 44) depends on at least the measured quantities (U, I, T), and the internal resistance (Ri), and / or the state of charge (SOC) of the energy storage device (16, 32, 42) and / or the open circuit voltage (U0) of the energy storage device (16, 32, 42) and / or the temperature (T) of the energy storage device (16, 32, 42) are determined. The method according to claim 7, characterized by this.

9. The method according to claim 1, characterized in that the ratio of the internal resistance (Ri) of the polarization resistance (Rpold) to the maximum load of the energy storage device (16, 32, 42) depending on at least one ambient quantity and / or state quantity (U0, SOC, T) of the energy storage device (16, 32, 42) is preserved.

10. The method according to claim 1, characterized in that for each type of energy storage device (16, 32, 42), the maximum value is preserved from the ratio of the internal resistance (Ri) of the polarization resistance (Rpold).

11. The method according to claim 1, characterized in that the determined ratio of the internal resistance (Ri) of the polarization resistance (Rpold) is stored in a correspondence table (72) and / or is represented in the form of a polynomial depending on at least one ambient quantity (U0, T) and / or state quantity (U0, T).

12. The prediction of the characteristic quantity (Up) is performed using a quantity (Udyn,τ) that reproduces the influence of the load history of the energy storage device (16, 32, 42), which is stored in advance depending on at least one state quantity (U0, T, SOC, Q, Ri), on the characteristic quantity (Up), and the quantity (Udyn,τ) is selected depending on the state quantity (U0, T, SOC, Qbat, Ri). The method according to claim 1, characterized in that.

13. The energy storage device (16, 32, 42) is applied with a load profile (50, 52) including at least one base load (51) and a peak load (53), and after a rest period until the rest voltage (U0) of the energy storage device (16, 32, 42) is reached, it is applied with the peak load (53), and the voltage transition (U) occurring at this time is evaluated to determine the quantity (Udyn,τ) that reproduces the influence of the load history of the energy storage device (16, 32, 42) on the characteristic quantity (Up). The method according to claim 12, characterized in that.

14. A sensor (34, 44) for capturing the measured quantities (U, I, T) is provided, and / or the sensor (34, 44) includes the state recognition (82), and / or the sensor (34, 44) is used to preserve the ratio between the polarization resistance (Rp ol) and the internal resistance (Ri), and / or the sensor (34, 44) is used to predict the characteristic quantity (Up). The method according to claim 8, characterized in that.

15. When the limit value (Ulimit) is reached due to the predicted characteristic quantity (Up), the interruption of a safety-related function and / or the output of a warning notification is started. The method according to claim 1, characterized in that.

Images