Predicted maximum electric charging power at a vehicle charging station

Abstract

The invention relates to a system for estimating the maximum achievable charging power of a charging station (1) of a charging park, comprising a respective device (7) for a vehicle (3), wherein a transmission unit of the device (7) transmits information about the current state of the charging park, determined by means of a sensor unit of the vehicle (3), to a computing unit (5), wherein the current state of the charging park includes at least a number of free and / or the number of occupied charging stations (1) in each case with respect to the total number of charging stations (1) in the charging park, and wherein the computing unit (5) analyzes the information determined by the sensor unit of the respective vehicle (3) and, based on this and on model data about the charging park, determines an estimate of the maximum charging power at an unoccupied charging station (1) of the charging park.

Description

[0001] The invention relates to a system for estimating an actually maximum achievable charging power of a charging station of a charging park with a plurality of charging stations for electrically charging battery-electric vehicles.

[0002] Public or semi-public charging parks for battery-electric vehicles generally specify a theoretical maximum achievable charging capacity per charging station / charging point within the charging park. In some cases, however, the actual achievable charging capacity at at least one charging point may be below this theoretical charging capacity. This could be due to the charging park actively throttling the charging capacity during parallel charging processes at the charging stations, particularly for vehicles with high charging capacities, or due to legal requirements regarding noise emissions (e.g., at night) or due to high grid utilization in this section of the power grid. The charging park may also reduce the nominally achievable charging capacity for specific individual charging points.

[0003] The most common cause of such a reduction is parallel charging processes of battery-electric vehicles with relatively high charging power, resulting in the total electrical power output of the charging park exceeding a maximum value. Typically, the charging stations used are controlled in such a way that the total available electrical power of the charging park or for a group of charging stations is distributed as equally or fairly as possible among the recipients. This ensures that multiple charging processes can take place in parallel and that no single vehicle alone consumes the full or by far the highest charging power.

[0004] In addition, rarer scenarios can arise that lead to a reduction in the charging power of a charging station. For example, the power of individual charging stations in residential areas can be deliberately reduced at certain times to ensure compliance with legally prescribed noise emission limits. Another possible reason could be the power grid load at the location. In particular, the widespread introduction of weather-dependent electrical power generation and clusters of charging battery-electric vehicles during peak times can lead to temporary overloads in the power grid. If the power flow exceeds a specified limit, the charging park may be required to throttle the entire output of electrical power.In the future, grid operators could also increasingly reserve the right to throttle electrical consumers such as wall boxes and heat pumps for a limited period of time, possibly at reduced grid charges, for example when current electricity prices are high.

[0005] However, for time-optimized and charging-stop-optimized route planning for a battery-electric vehicle, it is important to know the actual achievable charging power at one or more charging stations in a charging park. It is therefore desirable to be able to estimate the actual charging power currently or in the future with high accuracy, especially at charging stations that are not yet occupied.

[0006] State-of-the-art approaches focus primarily on the currently available charging power at each charging point in a charging park. However, these approaches only allow a vehicle manufacturer to successfully provide forecasts of available charging power to customers if the vehicle manufacturer also has access to data from the network or charging station operators.

[0007] In this context, EP 2 902 250 A2 relates to a method for charging an electrically operated vehicle at a charging point of a charging station having a plurality of charging points and connected to a power grid. Taking into account a maximum permissible charging cable current on the part of a charging cable and a maximum permissible limiting charging current on the part of the charging point of the charging station, a maximum permissible target charging current for charging the vehicle is determined; the maximum permissible target charging current is communicated to the vehicle by the charging station; an actual charging current, which is set via a charger provided in the vehicle, taking into account the maximum permissible target charging current and a maximum permissible vehicle charging current on the part of the vehicle, for the charging point, is measured at the charging station.wherein a charging current reservation is made at the charging station for the charging point in the amount of a maximum limit charging current that can be provided for the charging point by the hardware of the charging point and / or the maximum permissible target charging current offered to the vehicle, wherein a charging current difference is determined between the maximum permissible target charging current offered to the vehicle by the charging station or the maximum limit charging current that can be provided at the charging point on the one hand and the measured actual charging current on the other hand, and wherein the charging current reservation is released with regard to the determined charging current difference;

[0008] However, data from the electricity grid operator or the operator of a charging park with multiple charging points for battery-electric vehicles is not always available to a customer, especially to a driver of a battery-electric vehicle, before reaching the charging park. Estimates or measurements taken during the charging process itself naturally come too late for time-optimized and charging-stop-optimized route planning for an electric vehicle.

[0009] The object of the invention is therefore to be able to more reliably predict the actually achievable electrical charging power at a charging station of a charging park for battery-electric vehicles, even without receiving information about it from a power grid operator or an operator of the charging park itself.

[0010] The invention is based on the features of the independent claims. Advantageous developments and refinements are the subject of the dependent claims.

[0011] A first aspect of the invention relates to a system for estimating an actually maximum achievable charging power of an unoccupied charging station in a charging park with a plurality of charging stations for electrically charging battery-electric vehicles, comprising a computing unit and a respective device for one or more vehicles, wherein the respective device has a transmission unit and a sensor unit, wherein the transmission unit is designed to transmit information determined by the sensor unit of the vehicle to the computing unit, wherein the computing unit is designed to analyze the information determined by the sensor unit of the respective vehicle in order to determine a current state of the charging park, wherein the state comprises at least a number of free and / or the number of occupied charging stations, each in relation to the total number of charging stations in the charging park,and, based on the determined condition and based on model data on the charging park, to determine an estimate of the maximum charging power that can actually be achieved in the future at an unoccupied charging station in the charging park.

[0012] Implementing such a system solves the problem that neither a driver of a battery-electric vehicle nor a provider of charging route planners generally have a priori knowledge of the actual charging power at a charging station on a planned route or in a weekly profile. Instead of obtaining the data (exclusively) from a charging park operator, the relevant information is collected from vehicles in public transport (i.e., in particular, customer vehicles or fleet vehicles) in order to be able to provide a forecast of the electric charging power for a specific charging station.

[0013] In the above and in the following, the term charging station is understood logically and functionally rather than structurally, i.e. synonymous with a charging space and a charging point, i.e. that a charging station in the present context comprises an electrical connection for exactly one battery-electric vehicle.

[0014] A charging station is used to supply a battery-electric vehicle with electrical power to increase the vehicle's battery charge level. In this context, a battery-electric vehicle is understood to be a purely battery-electric vehicle, but also a hybrid vehicle that has at least one battery and another energy source.

[0015] The model data about the charging park is preferably stored in a memory of the computing unit and made available for retrieval. To obtain more complex and improved model data, a machine learning model can also be executed on the computing unit. This model has been defined for a specific charging park and can optionally be continuously improved to obtain current model data by executing the machine learning model, particularly if the information acquired by the sensor unit of a specific vehicle is used as input data for the machine learning model. Advantageously, an artificial neural network, ensembles of decision trees, or similar is used as the machine learning model.

[0016] For the basic functioning of the system, it is irrelevant whether a single vehicle with such a device or several vehicles, each with their own device, are used in the system to obtain external information about the status of a charging park. In principle, a single vehicle with such a device is sufficient, for example, to record the current status of a charging park when driving past it, in particular with regard to the current occupancy in terms of the number of vehicles already charging and the types of vehicles. However, it is advantageous to use many passing vehicles with such a device in order to obtain current information about the status of the charging park and, optionally, further information about prevailing boundary conditions and situation parameters. This allows not only the estimate to be updated, but also the model data or the machine learning model.

[0017] Furthermore, it doesn't matter where the interpretation of information from the sensor unit takes place. In other words, the processing unit can be a central component, but part of its function can also be outsourced to the device. For example, image analysis can be performed in the vehicle itself (with such a device), and the number and type of vehicles already charging in the charging park can be determined, as well as how many charging stations are occupied / still free.

[0018] A characteristic of the system is in any case the estimation of the maximum achievable electrical charging power of a charging station in a charging park with several charging stations, whereby the charging station in question is preferably not occupied at the time the information is acquired by the sensor unit, using external data, i.e. based on data from a sensor unit of a device in a vehicle. The external data therefore has one thing in common: it has been obtained from the ego perspective of a vehicle, but has not been transmitted by the charging park. Accordingly, the sensor unit is preferably one of a camera unit or a lidar unit, whereby sensor fusions from several sensors can also be used to obtain information about the status of a charging park. In particular when sensor fusions are used, other data can be included, for example radar data, data from ultrasonic sensors, and others.

[0019] The result of the computing unit, namely the estimate of the actual maximum achievable charging power at a specific charging station in a charging park, is preferably transmitted by the computing unit to a plurality of user vehicles. These user vehicles can use the received information to optimize route planning, to improve the estimated arrival time at the destination on a route with a charging stop at a charging park analyzed in this way, and for other purposes. From a technical perspective, it is irrelevant for the transmission of this result by the computing unit to user vehicles whether the receiving user vehicles themselves are equipped with such a device or not. Rather, the receipt of this estimate depends on whether the customer requests such information from the computing unit or whether it can technically be received in the corresponding receiving vehicle.

[0020] Preferably, the computing unit and / or device are designed to first detect whether at least one charging station is still unoccupied in the charging park, which is detected by the sensor unit of a respective vehicle. This is relevant, for example, because the number of still free charging stations influences the estimated maximum achievable electrical charging power at one of the free charging stations, and if all charging stations are fully occupied, a charging stop for another vehicle is at least not possible at the current time anyway.

[0021] According to an advantageous embodiment, the computing unit is a central computing unit and serves as a backend for the respective vehicle with the device.

[0022] According to a further advantageous embodiment, the state of the charging park comprises the number and type of currently charging vehicles in the charging park, wherein the computing unit is designed to analyze the information determined by the sensor unit of the respective device of a respective vehicle for the number and type of currently charging vehicles in the charging park, and to determine a currently available total charging power in the charging park by comparing the model data about the charging park and one of the determined number and type of currently charging vehicles and to estimate a remaining maximum achievable charging power of a still free charging station in the charging park.

[0023] Preferably, the computing unit is designed to determine the currently available total charging power in the charging park using a database, wherein the database contains a plurality of types of battery-electric vehicles and includes an assignment of each of the types to a preferred and / or maximum charging power. Thus, using the database, the computing unit can determine the current electrical power consumption at a charging station in the charging park by a respective vehicle already charging, and by summing these individual values, the total electrical power currently being drawn from the charging park can be determined.If this total charging power of the currently charging vehicles is compared with the model data for the charging park, the computing unit can determine whether the charging park is already at its full maximum electrical power or, if not, what electrical charging power could be available at a still free charging station.

[0024] According to a further advantageous embodiment, the model data on the charging park include empirical data values ​​on charging power delivered by one or more charging stations.

[0025] According to a further advantageous embodiment, the computing unit is designed to determine the model data by executing a machine learning model.

[0026] According to a further advantageous embodiment, the computing unit is designed to continuously update the machine learning model based on determined relationships between the actually received charging power and the state of the charging park.

[0027] Accordingly, the device is preferably designed to transmit a currently received charging power to the computing unit after the start of the charging process at the charging station for which the estimate was determined by the computing unit, if a machine learning model is already available, wherein the computing unit is designed to initially parameterize the machine learning model if not yet available and to update it if available.

[0028] According to a further advantageous embodiment, the device is designed to determine additional information about the charging park, in particular at least one of: geographical location of the charging park, individual current occupancy of charging stations with distinguishability of the charging stations, current temperature, current day of the week, current time, current solar radiation, current presence of rain, current wind, current weather data, current power grid utilization, in particular with regard to the geographical location of the charging park; wherein the computing unit is designed to use available additional information about the charging park to determine the model data.

[0029] Preferably, visual image processing or computer vision can be used to capture (display) information displayed at one or more charging stations in the charging park, which, for example, contains information on the current charging power, current battery level, previous charging time, remaining charging time, etc. Current information displays of the charging station(s) can therefore be captured.

[0030] Using the data, a model is created, for example using machine learning, or if already created, updated, that can predict the actual charging power at a charging park depending on contextual data such as location, time, day of the week, temperature, day, time of day, solar radiation, wind, grid utilization, current and predicted charging point utilization, etc., at the current time and for future points in time.

[0031] According to a further advantageous embodiment, the state of the charging park comprises a number of defective charging stations in relation to the total number of charging stations in the charging park, wherein the computing unit is designed to analyze the information determined by the sensor unit of the respective device of a respective vehicle for defective charging stations and, based thereon and based on model data about the charging park, to determine an estimate of the actually achievable maximum charging power at an unoccupied charging station in the charging park.

[0032] According to a further advantageous embodiment, the device is designed to output an estimate of the charging time in the charging park to a user, in particular the driver of the vehicle, depending on the estimate determined by the computing unit.

[0033] According to a further advantageous embodiment, the computing unit is designed to determine the obtained estimate for a plurality of user vehicles with or without the device.

[0034] A further aspect of the invention relates to a method for generating a machine learning model for estimating an actually maximum achievable charging power of a charging station in a charging park with a plurality of charging stations for electrically charging battery-electric vehicles, wherein a vehicle with a device comprising a transmission unit and a sensor unit with the sensor unit determines information about a current state of the charging park and transmits it to the computing unit, wherein the current state of the charging park comprises at least a number of free and / or the number of occupied charging stations, each in relation to the total number of charging stations in the charging park, wherein the determined information is analyzed by the computing unit and an actually obtained charging power of the charging station used is used by means of the device,to map the relationship between the information obtained and the charging power received in a machine learning model.

[0035] Further advantages, features, and details will become apparent from the following description, which – where appropriate with reference to the drawings – describes at least one embodiment in detail. Identical, similar, and / or functionally equivalent parts are provided with the same reference numerals.

[0036] They show: Fig. 1: A situation for applying a system for estimating an actually maximum achievable charging power according to an embodiment of the invention. Fig. 2: An application and update of a system for estimating an actual maximum achievable charging power in a first step according to an embodiment of the invention. Fig. 3: The application and update of the system based on Fig. 2 in a further step.

[0037] The representations in the figures are schematic and not to scale.

[0038] Fig. 1 shows a system for estimating the future actual maximum achievable charging power of a free charging station 1 in a charging park with a plurality of charging stations 1 for electrically charging battery-electric vehicles 3. The vehicle 3 has a device 7, which in turn has a sensor unit with a camera. This camera is used to first check the scene in the charging park and determine that three charging stations 1 are still free. Therefore, the actual maximum achievable charging power for charging the battery-electric vehicle 3 is estimated for one of the free charging stations 1, without obtaining data from the charging park itself. If all of the free charging stations 1 are of the same design, the estimate for one of the free charging stations 1 is sufficient. If different charging stations 1 are detected, a respective estimate is preferably made for all different charging stations 1.A transmission unit in the vehicle 3 is used to transmit the visual information about the current state of the charging park, captured by the sensor unit of the vehicle 3, to a cloud computing unit 5, which not only evaluates the image analysis of the current state of the charging park with regard to the occupancy of the individual charging stations 1, but also applies a machine learning model to estimate an expected maximum achievable charging power based on this information. This estimate of the maximum achievable charging power takes into account the fact that the theoretically specified maximum charging power of the charging stations 1 will not be achieved in many situations, for example, because the local power grid is reaching its limits and / or battery-electric vehicles with high electrical power are already charging in the charging park.In the central processing unit 5, which acts as the backend for each vehicle 3 equipped with such a device 7, a maximum charging power and average charging power, and optionally a charging curve, are read out for each detected vehicle type via a stored table. Thus, the following data is taken into account: - Number of vehicles during charging; - The respective vehicle type, as well as the corresponding maximum and average charging power or charging curve of the respective vehicle; - percentage of occupied and free charging stations; - percentage of defective charging stations;

[0039] The number of defective charging stations can be determined using image recognition if a corresponding indicator is displayed at each charging station, for example, using colored symbols or descriptive text. This means that the status of the charging park is known. Furthermore, the central processing unit 5 also takes current weather data, the current day of the week, and the current time of day into account. Model data on the charging park helps to determine an estimate of the maximum charging power that can actually be achieved in the future at an unoccupied charging station 1 in the charging park. The information collected above is the input of a machine learning model, which provides an estimate using analogous states and situation parameters from the past, such as the time of day, the day of the week, and the number and type of vehicles charging in the charging park.The resulting result regarding this estimated value can be used by the device 7 itself, for example, to explicitly communicate the estimated value to a driver of the vehicle. However, it can also optionally be transmitted to an entire fleet of user vehicles, preferably within a certain geographical area around the charging station, so that the user vehicles of this fleet can perform improved route planning.

[0040] In Fig. 2 is basically the situation of Fig. 1, but with the variant that the processing unit 5 is not a central processing unit, but a local processing unit of the vehicle 3. However, the execution of the processing unit 5 remains unchanged. Starting from a situation analogous to Fig. 1, vehicle 3 has now determined an estimated maximum power at which charging can be expected at a specific free charging station 1. Vehicle 3 now begins the charging process at the still free charging station 1. See further Fig. 3.

[0041] Fig. 3 shows how the vehicle 3 of the Fig. 2 reverse onto the space of the still free charging station 1 (see Fig. 2) has parked and is connected to the still free charging station 1. Device 7 of vehicle 3 can now compare the estimated value of the maximum achievable charging power with the actual maximum charging power value. Using the known "supervised learning" or "reinforcement learning" method, the parameters of the machine learning model can be adjusted, taking into account all recorded boundary conditions, such as the day of the week, time of day, and weather. The machine learning model is thus updated and can be passed on to other user vehicles. By specifying the required input data, a charging station 1 can now be used with known occupancy (number of vehicles, vehicle types, number of occupied charging spaces, number of free charging spaces) as well as context data such as time, location, temperature, charging point occupancy, etc.an estimated real charging power can be predicted within a short horizon, for example a few minutes, or at a specific time and day of the week. The prediction can be compared with current route planning for immediate entry, or it can be fed into route planning, for example for specific days of the week and times in the coming week. It can be assumed that the shorter the selected horizon, the more accurate the prediction will be. In order to keep the estimated real charging power up to date, and at the same time to expand the time horizon of the charging power prediction with current input variables, information on the number of vehicles, vehicle types, number of occupied charging stations, number of defective charging stations, number of free charging stations at a charging park as well as the local context information is continuously determined from the fleet of vehicles 3 and sent to a processing unit.Vehicles in the fleet that are not currently driving to a charging point but are driving directly past a charging park, such as vehicle 3, which is in the . Fig. 3 is drawn vertically and symbolically moves in the drawing plane from top to bottom. The predicted data can be sent to all vehicles in the fleet, for example via an interface of a central unit (which corresponds to the central processing unit 5 of the Fig. 1) can be made available to a respective fleet vehicle. This estimate can be used, in particular, for route optimization of battery-electric vehicles.

[0042] Although the invention has been illustrated and explained in detail by preferred embodiments, the invention is not limited by the disclosed examples, and other variations may be derived therefrom by those skilled in the art without departing from the scope of the invention. It is therefore clear that a multitude of variations exist. It is also clear that exemplary embodiments are truly only examples and should not be construed as limiting the scope, possible applications, or configuration of the invention in any way.Rather, the preceding description and the description of the figures enable the person skilled in the art to implement the exemplary embodiments in concrete terms, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or the arrangement of individual elements mentioned in an exemplary embodiment, without departing from the scope of protection defined by the claims and their legal equivalents, such as further explanations in the description. List of reference symbols 1 charging station 3 vehicles 5 Computing unit 7 Device QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] EP 2 902 250 A2

[0007]

Claims

[1] A system for estimating an actually maximum achievable charging power of an unoccupied charging station (1) of a charging park with a plurality of charging stations (1) for electrically charging battery-electric vehicles, comprising a computing unit (5) and a respective device (7) for one or more vehicles (3), wherein the respective device (7) has a transmission unit and a sensor unit, wherein the transmission unit is designed to transmit information determined by means of the sensor unit of the vehicle (3) to the computing unit (5), wherein the computing unit (5) is designed to analyze the information determined by the sensor unit of the respective vehicle (3) in order to determine a current state of the charging park, wherein the state comprises at least a number of free and / or a number of occupied charging stations (1), each in relation to the total number of charging stations (1) in the charging park,and based on the determined condition and based on model data about the charging park, to determine an estimate of the maximum charging power that can actually be achieved in the future at an unoccupied charging station (1) in the charging park. [2] System according to claim 1, wherein the computing unit (5) is a central computing unit (5) and serves as a backend for the respective vehicle (3) with the device (7). [3] System according to one of the preceding claims, wherein the state of the charging park comprises the number and type of currently charging vehicles (3) in the charging park, wherein the computing unit (5) is designed to analyze the information determined by the sensor unit of the respective device (7) of a respective vehicle (3) for the number and type of currently charging vehicles in the charging park, and to estimate a remaining maximum achievable charging power of a still free charging station (1) in the charging park by comparing the model data on the charging park and a currently available total charging power in the charging park determined as a function of the determined number and type of currently charging vehicles. [4] System according to one of the preceding claims, wherein the model data on the charging park comprise empirical values ​​on the charging power delivered by one or more charging stations (1). [5] System according to one of the preceding claims, wherein the computing unit (5) is designed to determine the model data by executing a machine learning model. [6] System according to claim 5, wherein the computing unit (5) is designed to continuously update the machine learning model based on determined relationships between the actually received charging power and the state of the charging park. [7] System according to one of the preceding claims, wherein the device (7) is designed to determine additional information about the charging park, in particular at least one of: geographical location of the charging park, individual current occupancy of charging stations (1) with distinguishability of the charging stations (1) and preferably their current information displays, current temperature, current day of the week, current time, current solar radiation, current presence of rain, current wind, current weather data, current power grid utilization, in particular with regard to the geographical location of the charging park; wherein the computing unit (5) is designed to use available additional information about the charging park to determine the model data. [8] System according to one of the preceding claims, wherein the state of the charging park comprises a number of defective charging stations (1) in relation to the total number of charging stations (1) in the charging park, wherein the computing unit (5) is designed to analyze the information determined by the sensor unit of the respective device (7) of a respective vehicle (3) for defective charging stations (1) and, based thereon and based on model data about the charging park, to determine an estimate of the actually achievable maximum charging power at an unoccupied charging station (1) of the charging park. [9] System according to one of the preceding claims, wherein the computing unit (5) is designed to determine the obtained estimate for a plurality of user vehicles with or without the device (7). [10] A method for generating a machine learning model for estimating an actually maximum achievable charging power of a charging station (1) of a charging park with a plurality of charging stations (1) for electrically charging battery-electric vehicles, wherein a vehicle (3) with a device (7) comprising a transmission unit and a sensor unit with the sensor unit determines information about the charging park and transmits it to the computing unit (5), wherein the determined information is analyzed by the computing unit (5) and a current state of the charging park is determined, wherein the current state of the charging park comprises at least a number of free and / or the number of occupied charging stations (1), each in relation to the total number of charging stations (1) in the charging park, wherein the state and a charging power actually obtained during charging of the charging station (1) used are used by means of the device,to map the relationship between the determined information or the state and the received charging power in a machine learning model.

Images