EP3863882A1

EP3863882A1 - Method and back end device for predictively controlling a charging process for an electric energy store of a motor vehicle

Info

Publication number: EP3863882A1
Application number: EP19795458.9A
Authority: EP
Inventors: Stefan Grubwinkler
Original assignee: Vitesco Technologies GmbH
Current assignee: Vitesco Technologies GmbH
Priority date: 2018-10-11
Filing date: 2019-10-09
Publication date: 2021-08-18
Also published as: CN112789193B; US20210387546A1; DE102018217454A1; WO2020074554A1; CN112789193A

Abstract

The invention relates to a method for predictively controlling (20) a charging process for an electric energy store (13) of a motor vehicle (12), wherein an energy exchange (14) between the energy store (13) and an electric energy source (15) is controlled by a charging device (11). According to the invention, a future time curve (25) of a non-energy requirement (N), resulting from a respective parking phase of the motor vehicle (12), is predicted, and the charge state (SOC) of the energy store (13) is kept below a threshold (21) by means of the charging device (11) regardless of the availability of a charge output of the energy source (15) if the predicted time curve (25) of the non-energy requirement (N) satisfies a specified rest criterion (22) for a specified subsequent time interval (23).

Description

description

Method and back-end device for predictive charge control for an electrical energy store in a motor vehicle

The invention relates to a method and a back-end device for predictive charge control for an electrical energy storage device of an electrically drivable motor vehicle. The charging control is intended to ensure the careful operation of the electrical energy store.

As an energy store, e.g. a battery with lithium-ion battery cells can be provided. The aging or wear of lithium-ion battery cells depends on the temperature and the state of charge (SOC) of the battery. For example, lithium-ion batteries should be stored below 10 degrees Celsius and at a medium (30-50%) or even low SOC (0-30%). With these storage conditions, the available battery capacity is reduced by less than 10% in 15 years, which is very advantageous.

However, with a reduced state of charge, the available range of a motor vehicle is also reduced. There are approaches that learn the destination of a trip on the basis of past trips and, for example, output the probable destination or the probable path to the destination (most probable path). Examples are US 20070150174 A1, US 8768616 B2, EP 2 042 833 A1, DE 10 2011 078 946 A1. These methods have the disadvantage that they only determine the destination in part depending on the first road segments traveled. However, this is too late to keep the energy store at a low SOC level (state of charge) because then it is no longer possible to adjust the state of charge by means of a charging device.

The invention is based on the problem of charging control for an electrical energy store of a motor vehicle. to sit back, the charging control should enable a gentle operation of the energy storage.

The task is solved by the subject matter of the independent claims. Advantageous embodiments of the invention are described by the dependent claims, the following description and the figures.

The invention provides a method for predictive charge control for an electrical energy store in an (electrically drivable) motor vehicle. The process assumes that an energy exchange of the

Energy storage with an electrical energy source can be controlled by a charging device. The energy source can be a power grid, e.g. B. act as a public power grid, i.e. an energy supply network. The charging device can have a connection via which the energy store of the motor vehicle can be connected to the energy source. An energy exchange or flow of energy between the power grid and the energy store can be controlled by the charging device. The charging device can be used as a stationary device, e.g. be configured as a charging station or as a charger for the motor vehicle. The question now is when and how much energy is charged into the energy store by the charging device or fed into the power grid from the energy store, i.e. which state of charge (SOC) should be set for a gentle operation of the energy storage.

According to the method, a future time course of a non-energy requirement is predicted or predicted. This non-energy requirement results from a respective parking phase or resting phase of the motor vehicle. The non-energy requirement can be, for example, the period during which the motor vehicle is coupled to the energy source via the charging device. In other words, the non-energy requirement always exists when the motor vehicle is not being driven. So the said time course of the non-energy requirement can indicate that Motor vehicle is coupled or connected to the charging device, that is, when the energy exchange by means of the charging device is possible.

In the method, it is now provided that a state of charge of the energy store is kept below a limit value by means of the charging device if the predicted time profile of the non-energy requirement for a predetermined next time interval satisfies a predetermined resting criterion. It is therefore taken into account when, in accordance with the time course of the non-energy requirement, no energy from the energy store is required, which is expressed in that the predicted one

Time course of the non-energy requirement for the predetermined next time interval fulfills the predetermined rest criterion. This time interval can mean, for example, that the rest criterion will be fulfilled for the next minute or the next hour, starting from the time in question. However, the time interval can also be 0 seconds, which means that only the time currently being considered is taken into account. The resting criterion can mean that the

Motor vehicle is not required, so no energy from the energy storage is necessary. For example, it can be specified as a resting criterion that the time course of the

Non-energy demand is above a predetermined threshold. The time interval under consideration can be selected in such a way that a charging time can be taken into account that may be necessary to charge the energy store, so that in the event that, according to the predicted time profile, the resting criterion is no longer met in a predetermined future time, then the Energy storage is also charged.

Thus, while the motor vehicle is coupled to the charging device, the future non-energy requirement can be determined, i.e. monitored or estimated or predicted, for how long the motor vehicle will be without energy requirement, i.e. remain unused and / or coupled to the charging device, based on a future course of time will stay. The charge level can then be kept below the limit for this time. It should be noted here that the state of charge can be kept below the limit regardless of the charging capacity available in the power grid. In other words, the state of charge does not need to be kept below the limit value because no charge power is currently available from the energy source and / or because at least one other motor vehicle is to be charged first and the total charge power is to be kept below a threshold value. Rather, the state of charge of the energy store can be kept below the limit value regardless of the availability of the charging power of the energy source. In the foreground is the lifespan extension of the energy storage. However, if the energy storage is to be integrated into a larger network (e.g. low-voltage network) consisting of several energy storage units with time-dependent electricity prices, the benefits can be weighed between the service life and the cost savings and the charging process can therefore be postponed. The limit value is preferably below 70% based on the fully charged state (100% state of charge), in particular below 60% below the fully charged state of charge. The limit value can also be set, for example, as a function of an ambient temperature of the energy store and / or the motor vehicle, which has already been described as advantageous at the beginning.

Thus, the motor vehicle, while it is connected to the charging device, is not unnecessarily burdened to the extent that the energy store has a state of charge which favors the wear of the energy store. Only when it is recognized from the predicted time course of the non-energy requirement that the resting criterion will be violated in the future or the non-energy requirement ends the other way round (and thus there is an energy requirement because the parking phase ends), can a charging process or energy exchange be started in order to store the energy for prepare the end of the parking phase or rest phase. Depending on when the resting criterion is violated in the future, a start of a charging process can be determined, for example. Here the available and / or predetermined charging power and / or the current state of charge can also be taken into account.

The invention also includes embodiments which result in additional advantages.

In one embodiment, said predicted time profile of the non-energy requirement for respective points in time of the time profile indicates that, starting from the respective point in time t, the motor vehicle does not require any electrical energy from the energy store and / or with the charging device and / or the energy source will remain coupled. This next period of time, calculated from the respective time t, can be, for example, in a range from 1 minute to 1 hour, for example within 1 minute to 30 minutes. The course of time at the respective times t indicates whether the motor vehicle will continue to require no electrical energy from the energy store, specifically for the next said predetermined period of time (for example 10 minutes) and / or whether the motor vehicle continues to operate for this period of time Charger will remain paired. For an exclusive consideration of the current point in time, the duration under consideration can also be 0 minutes. Thus, the time course of the non-energy requirement for different times indicates whether the motor vehicle will still be available for an energy exchange.

If the motor vehicle is connected to the charging device and the energy store is discharged to such an extent that the state of charge is below said limit value, the charging device can simply remain deactivated and / or the energy exchange can be omitted as long as the time course of the non-energy requirement fulfills the resting criterion. Here, it can of course be provided that in the event that the energy store is discharged below a predetermined minimum charge state, at least the minimum charge state is restored by an energy exchange. However, in the event that the motor vehicle is connected to the charging device and the state of charge is above the limit value, one embodiment provides that the energy store is discharged until the state of charge is below the current limit value. This ensures that even a motor vehicle with a fully charged energy store or an energy store whose state of charge is above the limit value is operated gently in the parking phase.

In one embodiment, said time profile is an indication of probability. The limit value below which the state of charge is maintained can be continuously adjusted as a function of the probability. If the course of time has a temporal fluctuation or change, a corresponding fluctuation or change can also be set for the limit value. “Continuous values” here means the difference to the continuous-time adjustment, which results from the adjustment anyway. As an alternative to the continuous-value adjustment, a graduated adjustment can be provided, that is to say a plurality of limit values or a group of limit values can be provided between which can be switched in stages. By taking the probability information into account when setting the limit value, the state of charge can be increased by recharging the energy store if the probability information for the non-energy requirement decreases, so that the driver is provided with an energy content by the increased state of charge, if the probability is realized or occurs that the motor vehicle is being used. In other words, the state of charge is kept at the limit value and not simply at any state of charge below the limit value.

The time course of the non-energy requirement can only indicate when the motor vehicle takes up the parking phase or the resting phase, that is, when it is coupled to the charging device. The said probability statement can also be provided here. If then the motor vehicle is not coupled to the charging device is, but is used or driving, it is advantageous to know how much energy the motor vehicle is then likely to need. Accordingly, a charging process can then be carried out or adapted before the end of the parking phase.

In one embodiment, a respective time profile of an energy requirement of the motor vehicle required from the energy store is also predicted. It is therefore estimated or stated for several different times or for several different time intervals how much energy the motor vehicle is likely to need in each case. This can be specified as a power requirement, for example. In the event that the time course of the non-energy requirement violates the resting criterion, that is to say the motor vehicle will end the parking phase, the charge state is set as a function of the predicted course of the energy requirement by means of the charging device. Because the violation of the resting criterion signals an impending use of the motor vehicle. Since this is now known in advance by means of the predicted time course of the non-energy requirement, a charging process can accordingly be initiated or started before the resting criterion is actually violated. How much energy is then required for the upcoming operating phase of the motor vehicle (ie the uninterrupted parking phase) can then be determined on the basis of the energy requirement in accordance with the predicted time profile of the energy requirement. The energy store therefore does not have to be fully charged (to 100% SOC), but the energy exchange from the energy source into the energy store can be limited, depending on the determined time course of the energy requirement. This has the advantage that unnecessarily much energy is not transferred to the energy store.

In one embodiment, the energy store is only charged to the extent that, according to the predicted time profile of the non-energy requirement, for the time interval that is from that point in time from which the resting criterion is violated, that is to say the parking phase has ended, to that point in time which the rest criterion is fulfilled again, i.e. the operating phase is over and the parking phase begins again, the state of charge of the energy store is in a time average in a range from 30% to 70% or in a range from 40% to 60%. During the operating phase, a state of charge will thus result which is on average over the said time interval in the said limits or range. This prevents an unnecessarily extreme state of charge (for example above 90%) from occurring on average. This protects the energy saver.

In one embodiment, however, a buffer value is added when the predicted time profile of the energy requirement is taken into account. The buffer value can be, for example, in a range from 1 kWh to 20 kWh. In this way, a forecast blur that can adhere to the time course of the energy requirement can be compensated for or taken into account. The state of charge is therefore higher than the expected predicted energy consumption. The buffer value is preferably specified in a user-specific and / or situation-specific manner and / or can be determined on the basis of the evaluation of the individual mobility behavior. In particular, the buffer value can be determined as a function of an individual mobility behavior of at least one in front of certain users. In this way, fluctuations or variance in a mobility behavior of the at least one user can be compensated for or taken into account.

In one embodiment, a prediction model is operated for the prediction of the respective time profile (the non-energy requirement and / or the energy requirement). This prediction model is configured or adapted to the motor vehicle. To configure the prediction model, at least some of the following trip data are recorded for at least one trip: time data (regarding, for example, days of the week and / or times of trips), weather data (regarding, for example, weather conditions for which the motor vehicle was used), route data (regarding starting locations and / or destinations and / or routes), consumption data (regarding, for example, a Average consumption and / or a driving style and / or a loading or loading of the motor vehicle), loading data for reloading operations (relating to loading operations at at least one other loading device). As a result, the prediction model can be adapted to an actually existing intended use of the motor vehicle. Taking time data into account makes it possible to directly predict the respective time profile. Taking weather data into account makes it possible to take into account conditional, weather-dependent use. In addition, the weather data can be used to determine an increased energy requirement for temperature control of the interior and the components of the motor vehicle (for example temperature control of the energy store itself) and can thus be taken into account in an energy requirement estimate. Taking route data into account makes it possible to recognize when a user knows a travel destination whether the user will be using the motor vehicle. It can also be used to determine the expected driving distances, which are the basis for the expected energy consumption. Taking consumption data into account makes it possible to predict energy consumption. Taking loading data into account makes it possible to plan for a reloading process at another loading device.

In one embodiment, said trip data is acquired by means of a vehicle-related acquisition. In other words, the use of the motor vehicle is taken into account, which can be carried out by several users. In addition or as an alternative to this, personal data can be recorded so that a specific user behavior of a particular user can be taken into account. Trip data for trips with at least one other person can also be transmitted to the user

Motor vehicle are detected. A charging control for several different motor vehicles that the user wishes to use can thus be provided for this user. The said travel data make it possible to configure the prediction model on the basis of historical or past driving processes that have been observed. In one embodiment, the acquisition of at least part of the trip data is carried out during at least one trip that is carried out with another motor vehicle, and the trip data acquired thereby are standardized by relating them to an average consumption and / or per predetermined distance unit (for example per Kilometers) and / or for different road classes (e.g. highway, country road, city) and / or road type (good

Road condition / poor road condition) can be determined. This means that a larger database is available. If a part of the trip data is determined here with a motor vehicle that has a different engine and / or a different weight, for example, normalization can ensure that an energy requirement for the motor vehicle that is connected to the charging device is correctly determined. Taking route units and / or road classes and / or road types into account makes it possible to determine, in the case of a planned route of the motor vehicle, what energy it is likely to have for this. The planned route can e.g. can be received in the form of navigation data from a navigation device.

In one embodiment, the predicted time profile is additionally adapted by the prediction model as a function of at least some of the following situation data. A mobility matrix can be taken into account that describes a mobility behavior or movement behavior of at least one user of the motor vehicle, that is to say indicates at what times the respective user is going from which starting point to which destination. Booking data for the motor vehicle can be taken into account, which indicate when at least one user has announced or ordered the use of the motor vehicle. Traffic data from road traffic can be taken into account, which can indicate whether a certain route can be used and / or the expected average speed at which the route can be used. Weather forecast data can be taken into account, which can indicate which weather or which weather is likely to prevail. This is particularly advantageous with the Combination that also determines the weather, the weather or the weather in which the motor vehicle is used and / or how high the additional energy requirement for conditioning the interior is. Personal activity data of at least one user of the motor vehicle can be determined, such activity data being determined, for example, by means of a mobile terminal (for example smartphone and / or tablet PC and / or smart watch) or generally a so-called “wearable device” The activity data can describe a current activity and / or a planned activity of the respective user, wherein the activity data can comprise at least one of the following data types: calendar data (for example at least one appointment of the user), alarm clock data (for example times for getting up and / or going to bed), movement data (for example a movement towards or away from the motor vehicle), a sewing indication of the motor vehicle (how far the user is from the motor vehicle), sleep phase information (ie when the user is sleeping) so additional In addition to the described historical trip data, such information is also taken into account that describes the current situation and / or a future and / or planned situation. So you can react to a change in your daily rhythm. This is also possible without interaction with the user.

In one embodiment, data from a respective wearable of the at least one user are taken into account in order to determine the intended vehicle non-use of the motor vehicle with respect to at least one predetermined user. Such a wearable can be a mobile end device, such as a smartphone or a tablet PC or a smartwatch, or it can be, for example, a device worn on the body, such as a fitness bracelet, or a piece of jewelry with a radio device. The wearable can be used to record a location and / or movement behavior of the user and to conclude that the motor vehicle is about to be used or not used (for example on the basis of a distance and / or a movement with respect to the force). vehicle). For this purpose, the said data can include geoposition data and / or appointment data. The data can also be the said activity data in order to adapt the respective course of time.

In one embodiment, in the event that a recharging process is detected while the motor vehicle is in use, the prediction model is corrected on the basis of the detected recharging process. The reloading process is in particular an unplanned or unforeseen or unpredictable reloading process. So there was a forecast error here. By correcting the prediction model, this prediction error can be reduced for further predictions of time courses. The prediction model can be implemented, for example, on the basis of a machine learning method, that is, for example, on the basis of a neural network and / or a decision tree and / or a regression model and / or a deep learning method ( Deep learning process). Additionally or alternatively, a statistical method (for example a Markov chain and / or a probabilistic network) can be used. With these methods, it is possible to adapt the prediction model as a function of a detected prediction error when a prediction error is detected, such as when detecting a reloading process.

In one embodiment, user input is received via a user interface (for example an Internet portal and / or a user program (for example application software for a mobile terminal). The user input announces a planned use of the motor vehicle. In other words, The user then explicitly specifies when he would like to use the motor vehicle. The predicted time profile of the non-energy requirement is then corrected in accordance with the user input. This advantageously allows a user to announce an exceptional situation and it is then ensured that the motor vehicle's energy store has a corresponding one Has state of charge. Depending on the user input, the time course of the energy requirement can also be adapted if the user input also defines, for example, a destination and / or a route.

In one embodiment, the prediction model is operated by a backend device. Such a back-end device is a device external to the vehicle, which in particular can also be operated at a distance from the charging device. It can be, for example, a computer or a group of computers. The back-end device can be implemented, for example, as an Internet server or cloud device. The implementation of the method by a back-end device has the advantage that data sources are used that are available outside the motor vehicle.

Said backend device also belongs to the invention. This has a computing device which is set up to carry out an embodiment of the method according to the invention. The computing device can be implemented on the basis of at least one microprocessor. The method can be implemented, for example, as a program code for the computing device. The program code can have program instructions which, when executed by the computing device, carry out the embodiments of the method according to the invention. The program code can be stored in a data memory or a non-volatile storage medium of the computing device. The backend device can be coupled to the charging device, for example via a communication link, in order to control the charging device. The communication device can be implemented on the basis of an Internet connection. The communication connection can include a radio-based connection, such as can be implemented, for example, on the basis of a mobile radio connection and / or WLAN connection (WLAN - Wireless Local Area Network). In the following, exemplary embodiments of the invention are described. This shows:

1 shows a schematic representation of a system architecture with an embodiment of the back-end device according to the invention;

Fig. 2 is a diagram with a predicted time course of a

Energy requirements;

Fig. 3 is a diagram with a predicted time course of a

Non-energy needs; and

4 shows a flow chart diagram of an embodiment of the method according to the invention.

The exemplary embodiment explained below is a preferred embodiment of the invention. In the exemplary embodiment, the described components of the embodiment each represent individual features of the invention that are to be considered independently of one another, which further develop the invention independently of one another and are therefore also to be regarded individually or in a combination other than the one shown as part of the invention. Furthermore, the described embodiment can also be supplemented by further features of the invention that have already been described.

In the figures, elements with the same function are each provided with the same reference symbols.

1 shows a system architecture 10, by means of which at least one charging device 11 can be controlled. Several charging devices can also be controlled in the manner described below. Only one charging device 11 is described below by way of example. The charging device 11 can be, for example, a charging station or a charging station. The Charging device 11 can also be integrated into motor vehicle 12.

In the case of a motor vehicle 12, the charging device 11 can be used to recharge its electrical energy store 13 by means of an energy exchange 14. The energy store 13 can be, for example, a high-voltage battery, that is to say a battery that can generate a DC voltage of more than 60 V. The energy store 13 can be a so-called traction battery of the motor vehicle 12, by means of which an electric drive unit of the motor vehicle 12 can be operated. For the energy exchange 14, the charging device 11 can be connected to an electrical energy source 15. The energy source 15 can be a public electricity network and / or an installation for regenerative energies, for example a photovoltaic installation. In general, the charging device 11 is therefore connected to an electrical energy source.

The energy exchange 14 can be controlled in the charging device 11 by control signals 16, which can be generated by a Steuerervor device 17. The control device 17 can in particular be designed as a back-end device 18, that is to say for example as an Internet server or cloud device for the Internet. The control signals 16 can be transmitted from the back-end device 18 to the charging device 11 via a communication link 19. The communication link 19 can be based, for example, on an internet connection and / or a radio link.

A charging control 20 can be implemented by the back-end device 18, which is used for the energy store 13 of the

Motor vehicle 12 provides a charging strategy in which it can also be provided that the state of charge of the energy store 13 is maintained if the motor vehicle 12 is not used within a predetermined next time period. It can hereby be provided that the state of charge is kept below a predetermined limit value 21. This For a storage type of the energy storage device 13, the limit value indicates that the energy storage device 13 is operated more gently if the state of charge is below the limiting value 21, as long as the motor vehicle 12 is not used, that is to say no energy is drawn from the energy storage device 13. The limit value 21 can be dependent on a storage technology of the energy store 13. One possible storage technology is lithium-ion storage technology. The limit value 21 can be taken, for example, from specialist publications.

The state of charge SOC can be set here by the energy exchange 14. As long as a rest criterion 22 is fulfilled, which indicates that the motor vehicle 12 will not be used within a next time interval 23, the charge state is kept below the limit value 21. In order to be able to recognize or predict whether the motor vehicle 12 will be used for a future time period, in particular for the next future time interval 23, a predicted time profile 24 for an energy requirement E and a predicted time profile 25 for a non-energy requirement N can be set in the back-end device 18 be taken as a basis.

FIG. 2 shows an example of a time profile 24 for an energy requirement E. The energy requirement E in the unit kilowatt hours (kWh) is shown over the time t (indicated in hours h of the day, that is to say from midnight to midnight). This also results in a power requirement of the motor vehicle 12. The energy requirement E indicates the energy required from the energy store 13. It can be the energy that the motor vehicle 12 requires when driving or driving.

FIG. 3 shows an example of a time profile 25 for the non-energy requirement N. It can be a statement of probability, which can be expressed, for example, in percent. The information is again over the time t in hours h of the day. The non-energy requirement N can indicate the probability with which the motor vehicle 12 is in a parking phase 26 (see FIG. 1) and with the charging device 11 or Energy source 15 is coupled, so that the charging control 20 can be carried out or implemented. If, on the other hand, there is no parking phase 26, but an operating phase, the motor vehicle 12 is decoupled from the charging device 11 or the energy source 15, so that no energy exchange 14 for setting the state of charge SOC is possible.

FIG. 4 illustrates a method that can be carried out in the backend device 18 by a computing device 27 (see FIG. 1) in order to implement the charging control 20. In a step S10, the future time profile 25 of the non-energy requirement N and the time profile 24 of the non-energy requirement N can be predicted. In a step Sil, the state of charge SOC of the energy store 13 can then be kept below the limit value 21 independently of the availability of a charging power of the energy source 15 by means of the charging device 11 if the predicted time profile 25 of the non-energy requirement N for a predetermined next time interval 23 is the resting criterion 22 Fulfills. The resting criterion can say, for example, that the time profile 25 for the non-energy requirement N must be above a predetermined threshold value 28 so that it is assumed that the motor vehicle 12 is actually coupled to the charging device 11 for the respective time t. It can also be provided that it is assumed that the time profile 25 for the future time interval 23 must be above the threshold value 28 from the current point in time.

3 illustrates how a time interval 32 arises between a point in time 30 at which the threshold value 28 is undershot by the time profile 25 and a point in time 31 at which the time profile 25 again exceeds the threshold value 28 the charging control 20 must assume that the motor vehicle 12 is in an operating phase and thus can only extract energy from the energy store 13. Provision can be made to set the state of charge SOC before the time 30 by the energy exchange 14 to such an extent that according to the time Course 24 for the energy requirement E (FIG. 2) gives an average state of charge SOC which is in a range from 30% to 70% state of charge, in particular in a range from 40% to 60%. For this purpose, the energy requirement E for the time interval 32 can be determined from the time profile 24.

1 further illustrates how the time profiles 24, 25 can be predicted.

Provisional time profiles 24 ', 25' can first be determined by means of a prediction model 33. The prediction model 33 can be configured on the basis of historical trip data 34. The at least one motor vehicle 35 can be the motor vehicle 12, but it can also be one or more other motor vehicles. Each of the motor vehicles 35 can also have an energy store 36, so that charging processes also result for the at least one motor vehicle 35. The travel data 34 can be acquired from the at least one motor vehicle 35 by means of a data acquisition 37 which, for example, can acquire the travel data 34 on the basis of communication via a respective communication link 38 to the at least one motor vehicle 35. In addition, charging data 40 for charging processes of the at least one motor vehicle 35 can also be recorded by data acquisition 39 as trip data. Such loading data 40 can be received from the respective motor vehicle 35 and / or from a charging station. For this purpose, a corresponding communication link 41 can be provided. The communication connections 38, 41 can each comprise, for example, an Internet connection and / or a mobile radio connection and / or a WLAN radio connection.

On the basis of the recorded travel data 34, the prediction model 33 can be formed, for example, as a statistical model and / or as a model based on a machine learning method. The provisional time profiles 24 ', 25' can then be compared with actually observed time profiles, from which error data 42 can be generated which can be used for Correcting or improving the prediction model 33 can be used.

In addition to the historical journey data, current situation data 43 can also be taken into account, on the basis of which a current situation of motor vehicle 12 can be determined. From this, a respective correction 44, 45 for the provisional course of time 24 ', 25' can be carried out, which then results in the final estimated or predicted course of time 24, 25 in each case. For example, weather forecast data 46 from a weather station 47 and / or traffic data 48 from a traffic monitoring 49 can be used as situation data 43. Weather data and / or traffic data are preferably used to determine an increased energy requirement due to (e.g. low or high) outside temperatures and / or due to traffic congestion. Activity data 50, which can describe an activity of a user of the motor vehicle 12, can also be used. These can be received, for example, from a mobile terminal 51, which can be, for example, a smartphone and / or a tablet PC and / or a smart watch of the user. A mobility matrix 52 and / or booking data 53 from a corresponding data source 54 can also be used. A user interface 55 can also be provided, for example by means of the mobile terminal 51, through which a user input 56 can be received, through which the user can expressly indicate when he wants to use the motor vehicle 12. A correction 44, 45 for the

Time profiles 24 ', 25' are derived or determined.

The need to keep the state of charge low when a vehicle battery is not in use to avoid wear and tear shows that the conditions in non-driving operation are important for ensuring battery life, especially with smaller batteries. To do this, it is necessary to know well in advance (several hours or days) before the start of the journey when how much energy is required or at least to know when no energy is required. Current charging strategies of the prior art do not take into account the life of the batteries when choosing the time for the charging process. Numerous electric vehicles are charged immediately after inserting the charging plug. There are also approaches in which the user can individually determine when the vehicle can be charged. There are also smart grid solutions in which the charging time is charged depending on the electricity supply or price (controlled charging)

During the charging process, the charging current is selected in such a way that the life of the energy store during the charging process is impaired as little as possible.

However, due to the limited ranges of electric vehicles, users have the desire to have the highest possible energy content at the start of the journey (to avoid stopping or due to the limited availability of the public charging infrastructure).

It would therefore not be beneficial for the life of the battery storage to charge the energy storage after the end of the journey and to park the vehicle with a high SOC, so that the driver has a full energy storage at the start of the journey. It is rather necessary to fully charge the energy storage device shortly before the trip if possible and to keep the storage device at a low SOC level for as long as possible.

Determining the energy requirement well before the start of the journey is a problem. Research shows that only less than 25% of drivers use a navigation system. The drivers are therefore often not ready to enter the destination of the trip (especially in known areas). For a charging strategy to testify to electric driving, the driver would have to be ready to enter the next destination at the end of the journey.

It therefore becomes a method for learning the start and destination locations of the user on the basis of recorded travel data. These procedures are for the extended life charging problem Suitable for vehicle batteries. However, preferred starting and destination locations can often only be reliably determined with private individuals. In the commercial sector (e.g. craftsmen, representatives, company vehicles, etc.) such repeated and preferred starting and destination locations cannot be determined, so that predictive charging control should be implemented differently for these user groups.

Approaches are being implemented that predict when no energy is needed for the vehicle.

The following disadvantages can be eliminated:

Energy requirements for the entire route are only known after entering the navigation data

The energy requirement of the entire route is only known after the most probable route has been identified

User identification is required to identify the most likely overall route

Controlled recharging of the energy store (by user, smart grid) without taking battery life into account Non-existent strategy for keeping the energy store at a low SOC level (below a predetermined limit value).

The server-based system system architecture 10 described uses various data sources, processes the raw data and uses machine learning methods to predict the following variables (see FIG. 2):

Probable time course of the energy demand Time course of the probability of a non-energy demand (continuous for the next hours / days)

Two prediction variants are possible, a driver-specific variant (this primarily uses personal driver data) and a vehicle-specific variant (particularly suitable for fleet applications).

Based on the predicted values, an intelligent control of the charging strategy is used. This is wireless or wired communication connection and interface with the control unit of the charger of the electric vehicle is necessary. The individual components of the overall system are described below.

The following application in the form of a charging strategy for lithium-ion batteries is therefore possible. The charging strategy uses the predicted course of the energy requirement in order to have the energy storage fully charged at the expected start time. The point in time at which the charging process is to be started, depending on the current SOC of the battery, is calculated so that the energy store is fully charged. It must be known how much power the electric vehicle can be charged with (e.g. 3kW household socket, wall-charging station with a truck or 22 kW).

If the expected journey is very short, the energy store is only fully charged so that the destination can be reached with an average SOC value (e.g. 50%).

The probability of the non-energy requirement is used to determine the SOC level during the parking process / vehicle standstill. If there is a higher probability of non-energy demand (e.g. at night), the SOC level is left at a low level in order to minimize the lifespan of the energy store. If the probability of the non-energy requirement decreases, the SOC level is increased by recharging the energy store, so that the driver has a certain amount of energy at his disposal should he still want to start unexpectedly.

With a smart grid connection of the vehicle, it is also possible to discharge the energy storage device after the end of the journey (e.g. in decentralized home storage systems) in order to lower the SOC level if the energy storage unit is very full after the end of the journey and for the next few hours no ride is expected. Since the predicted values have a certain uncertainty about the prognosis, a certain safety buffer is taken into account in the loading strategy. The vehicle is fully charged a certain period of time before the start of the journey and the SOC level is higher than the expected predicted energy consumption.

Each time a vehicle driver drives, trip data 34, such as the start of the trip, the outside temperature (by weather data), the duration of the trip, the travel distance and the energy consumption of the vehicle (electrical or conventional) can be recorded. This data can be recorded on the basis of GPS data and available OBD data in conventional vehicles. In the case of electric vehicles, there is an interface to the vehicle's communication network so that the data can be recorded. The data is transmitted to a backend via wireless communication. The data is recorded personally (for private vehicles) or vehicle-related (for fleet vehicles).

The recorded energy consumption of the vehicle is standardized so that the energy consumption of different vehicles can be compared. For this purpose, the energy consumption per unit of distance can be simply divided by the average consumption of the vehicle. Alternatively, the route-dependent average energy consumption depending on the road class or the road type can be used for standardization. To do this, digital map data must be used to determine which sections of the route have been covered with which road class or type. This can happen while driving or in the backend with the recorded driving trajectories.

In addition, all charging processes are recorded (duration, amount of energy, time).

In addition to the recorded sources, additional data sources (link to backend necessary) can be used to increase accuracy. Examples:

Mobility behavior: Origin - Destination Matrix / prediction of the probable destinations by evaluating the individual driver behavior (with driver-specific recording)

Appointment calendar of the driver (driver-specific entry) Booking system of the fleet (vehicle-specific entry)

Traffic forecast / traffic patterns (to correct the predicted energy requirement)

Weather conditions (to correct the predicted energy requirement)

Data from wearables / smartphones (to correct the likelihood of no need for driver-specific recording):

o Analysis of the sleep phases

o Alarm clock parameterization

o Current distance to the vehicle

For the prediction model 33, it is statistically recorded when which energy consumption is required. The statistical evaluation takes place every hour and every weekday. The use of statistical methods (Markov chains, probabilistic networks, to name just a few examples) and machine learning processes (neural networks, decision trees, regression models, deep learning processes, to name just a few examples) result in the following Predicted values:

Energy requirements

Probability for no need of the electric vehicle (non-energy requirement)

The course of the values over time for a parameterizable prediction horizon (e.g. 24h) is calculated. The prediction is based on the recorded data at the time of vehicle use and the recorded energy consumption values. Prediction is therefore possible without entering user data or entering navigation data.

In a second step, the prediction is corrected depending on the additional data sources available. For this purpose, the energy consumption depends on the vehicle used expected weather conditions and the expected traffic conditions have been corrected.

In the case of driver-specific recording, the probability of no need is corrected depending on the user-specific data available (e.g. by knowledge of smartphone data such as the alarm clock, by knowledge of the appointment calendar or by evaluation of personal mobility behavior).

The prediction model is improved by evaluating recharging on the go (on non-preferred charging stations). For example, it analyzes whether the

Reloading would have been avoided if the memory had already been fully loaded at the start of the journey. If an inaccurate forecast is the cause of the reloading process, the forecast model is adjusted accordingly. For example, methods of reinforcement learning are used.

The user receives a display of the predicted values through a user interface (user interface 55, for example in the form of a web portal, SmartPhone and / or application software). The system predicts the energy requirement and the probability of a non-energy requirement regardless of the user input. However, in exceptional situations (e.g. when traveling on vacation at 3:00 a.m.), the user has the option of specifying a required start of the journey to ensure that the energy storage is sufficiently charged in these exceptional situations.

Since it is a learning system and the forecast is increasingly adapted to user behavior, user inputs are useful in the learning phase.

This results in the following options:

Prediction of a probability for the non-energy need to determine the lowest possible

SOC level during the parking of electric vehicles, Strategy for charging the energy storage as late as possible before starting the journey,

Charge management to extend the life of the energy storage without the need for user data or user specifications,

Linking charge management with smartphone or

Wearables to determine if a vehicle is needed

The approaches described can also be used in the following areas:

Avoiding battery aging,

Preconditioning of air conditioning,

Fleet management for fleet vehicles,

Smart grid applications: consideration of energy requirements of electric vehicles for decentralized storage; Feeding back electric vehicles into decentralized house storage

Overall, the example shows how the invention provides the prediction of a probability that an electric vehicle will not be required to extend the battery life of the electric vehicle.

Claims

1. A method for predictive charging control (20) for an electrical energy store (13) of a motor vehicle (12), wherein an energy exchange (14) of the energy store (13) with an electrical energy source (15) is controlled by a charging device (11), that is, that

shows that a future course of time (25) of a non-energy requirement (N), which results from a respective parking phase of the motor vehicle (12), is predicted and a state of charge (SOC) of the energy store (13) by means of the charging device (11) below one Limit value (21) is maintained if the predicted time profile (25) of the non-energy requirement (N) for a predetermined next time interval (23) meets a predetermined resting criterion (22).

2. The method according to claim 1, wherein the predicted time profile (25) of the non-energy requirement (N) indicates for respective points in time that the motor vehicle (12), starting from the respective point in time, has no electrical energy from the energy store (13 ) is required and / or will remain coupled to the charging device (11) and / or the energy source (15).

3. The method according to any one of the preceding claims, wherein in the event that the state of charge (SOC) is above the current limit value (21), the energy store (13) is discharged.

4. The method according to any one of the preceding claims, wherein the time course (25) is a probability statement and the limit value (21) is adjusted as a function of the probability statement values-continuously or in stages.

5. The method according to any one of the preceding claims, wherein a time course (24) of an from the energy store (13) be required energy requirement (E) of the motor vehicle (12) is predicted and in the event that the time course (25) of the non-energy requirement (N ) violates the resting criterion (22) and thereby one impending use of the motor vehicle (12) is signaled, the state of charge (SOC) is set as a function of the predicted time profile (24) of the energy requirement (E) by means of the charging device (11).

6. The method according to claim 5, wherein the energy store (13) is only charged to such an extent that for a time interval (32) which differs from a point in time (30) from which the resting criterion (22) is violated to a point in time (31), for which the resting criterion (22) is fulfilled again, the state of charge (SOC) lies in the time average in a range from 30 percent to 70 percent or in a range from 40 percent to 60 percent.

7. The method according to claim 5 or 6, wherein when taking into account the predicted time profile (24) of the energy requirement (E), a buffer value is added, which is determined as a function of an individual mobility behavior of at least one predetermined user.

8. The method according to any one of the preceding claims, wherein a prediction model (33) is operated for the prediction of the respective time profile (24, 25) and at least some of the following journey data (34) are recorded for configuring the prediction model (33) for at least one journey : Time data, weather data, route data, consumption data, loading data (40) for reloading processes.

9. The method according to claim 8, wherein the journey data (34) are determined by means of a vehicle-related detection and / or a person-related detection.

10. The method according to claim 8 or 9, wherein the detection of at least part of the trip data (34) takes place during at least one trip that is carried out with another motor vehicle (35), and the trip data (34) detected thereby are standardized by: are related to an average consumption and / or are determined per predetermined route unit and / or road class and / or road type.

11. The method according to any one of claims 8 to 10, wherein the respective predicted time profile (24, 25) is adapted as a function of at least some of the following situation data (43):

- a mobility matrix (52) of at least one user of the motor vehicle (12),

- Booking data (53) for the motor vehicle (12),

- traffic data (48) from road traffic,

- weather forecast data (46),

- personal activity data (50) of the at least one

Users who describe a current activity and / or a planned activity, the activity data comprising at least one of the following data types: calendar data, alarm clock data, movement data, a sewing information on the

Motor vehicle, sleep phase information.

12. The method according to any one of claims 8 to 11, wherein in the event that a recharging process is detected during use of the motor vehicle (12), the prediction model (33) is corrected on the basis of the detected recharging process.

13. The method according to any one of the preceding claims, wherein data from a respective wearable of the at least one user are taken into account to determine the intended vehicle disuse of the motor vehicle with respect to at least one predetermined user.

14. The method according to any one of the preceding claims, wherein a user input (56), through which a planned use of the motor vehicle (12) is announced, is received via a user interface (55) and the predicted time profile (25) of the non-energy requirement according to the user input ( 56) is corrected.

15. Backend device (18) with a computing device (27), characterized in that the re- Cheneinrichtung (27) is set up to carry out a method according to any one of the preceding claims.