Method for predicting the energy demand of an electrical consumer component in a motor vehicle

Abstract

The present invention relates to a method for predicting the energy demand of an electrical consumer component in a motor vehicle (1). According to the invention, a functional model (6) is generated and / or trained in advance for the electrical consumer component (3) from data of a bus system (2) to which the electrical consumer component (3) is linked, using artificial intelligence, wherein the functional model (6) is subsequently used in the motor vehicle (1) to forecast values ​​for the energy requirement of the consumer component (3) from the data of the bus system (2).

Description

[0001] The invention relates to a method for predicting the energy demand of an electrical consumer component in a motor vehicle. The invention also relates to a motor vehicle in which the method is used.

[0002] DE 10 2021 006 280 A1 discloses a method for predicting the energy consumption of a motor vehicle. This method uses artificial intelligence to determine a functional model for predicting the vehicle's energy consumption from data from a motor vehicle's bus system. This functional model is then used to predict the vehicle's energy consumption from the bus system data.

[0003] The object of the present invention is to provide an improved or at least different method for predicting the energy demand of an electrical consumer component in a motor vehicle.

[0004] This object is achieved according to the invention by the subject matter of the independent claims. Advantageous embodiments are the subject matter of the dependent claim.

[0005] The core idea of ​​the invention is therefore to use a functional model generated and / or trained by means of artificial intelligence to predict the energy demand of an electrical consumer component in a motor vehicle, wherein the functional model is generated and / or trained using data from a bus system of the motor vehicle to which the consumer component in the motor vehicle is connected. This functional model is used during operation of the motor vehicle to forecast values ​​for the energy demand of the consumer component for given points in time based on the bus data available during operation. The data provided for generating and / or training the functional model can, for example, originate from a test vehicle and / or be determined during the development of the motor vehicle.The core idea of ​​the invention thus makes it possible to dispense with, or at least reduce, the use of measurement hardware, such as sensors, and / or evaluation hardware to determine the energy consumption of the consumer component during further testing and / or in corresponding series vehicles. This not only provides cost advantages, but also reduces the effort required to determine / predict the energy consumption of the consumer component during testing, further development, and / or maintenance of the motor vehicle.

[0006] The invention therefore proposes a method for predicting the energy demand of an electrical consumer component in a motor vehicle, in which method the functional model is first generated and / or trained and the thus generated and / or trained functional model is then used in the motor vehicle to predict the energy consumption of the consumer component. The functional model is generated and / or trained using artificial intelligence. For this purpose, a time-dependent bus data history of a bus system of the motor vehicle to which the consumer component is connected and consumption values ​​of the energy demand of the consumer component associated with the bus data history at given times are provided to the artificial intelligence as data pairs. The artificial intelligence generates and / or trains the functional model using the data pairs.The functional model thus generated and / or trained is used in the motor vehicle to predict the energy consumption of the consumer component. This is achieved by supplying / providing the functional model with bus data from the motor vehicle during operation. Using the functional model, the bus data is used to forecast corresponding values ​​of the energy demand of the electrical consumer component for at least two points in time or for a defined period, in particular repeatedly. The thus forecast values ​​are stored in a memory system and / or displayed to the user via a display. The display can be near real-time or delayed.

[0007] The stored energy consumption values ​​can, for example, be used to approximate with sufficient forecast quality if, for example, energy consumption is relevant over a longer period of time.

[0008] The display provides a real-time visualization of the predicted energy demand of the consumer component. This allows a user, for example, during vehicle testing and / or development, to take energy demand values ​​that deviate from a predetermined pattern into account and / or to conduct further investigations of the consumer component. Likewise, the consumption behavior of the consumer component during a journey and / or while stationary can be made visible and transparent to the user. Furthermore, indirect determination is possible without the use of appropriate measuring technology.

[0009] In advantageous variants, the bus data is simultaneously interpreted, i.e., decoded, via a so-called "library." The decoded data is advantageously fed to the artificial intelligence for generating and / or training the functional model and / or fed to the functional model during operation.

[0010] The generation and / or training of the functional model using artificial intelligence can be done in any way. For example, it can be done using a neural network and / or a regression model and / or an alternative machine learning model.

[0011] When generating and / or training the functional model, a maximum permissible inaccuracy for deviations between the predicted energy demand values ​​and the provided data pairs is preferably specified. Thus, with sufficiently high accuracy, the effort required to generate and / or train the functional model can be kept within limits. This results in further cost advantages while simultaneously ensuring sufficient accuracy of the predicted energy demand values.

[0012] It is understood that in addition to the method, a motor vehicle is also included in the scope of this invention, in which the functional model is used to predict the energy demand of the consumer component.

[0013] The motor vehicle thus has an electrical consumer component and a bus system, with the consumer component being connected to the bus system. The motor vehicle also has a control device which, during operation, receives bus data from the bus system. The functional model is stored in the control device. Alternatively or additionally, the functional model can be made available to the control device, for example, via a data connection. The control device is designed such that it uses the functional model to forecast values ​​of the energy demand of the consumer component as described.

[0014] It is understood that the functional model can also be designed for two or more such electrical consumer components. This means, in particular, that the functional model is generated and / or trained using artificial intelligence in such a way that corresponding energy demand values ​​are predicted for at least two consumer components. Subsequently, corresponding energy demand values ​​are predicted for at least two electrical consumer components in the motor vehicle from the bus data using the generated and / or trained functional model.

[0015] Likewise, the method can use such functional models for at least two different electrical consumer components of the motor vehicle. This means that these functional models are each associated with a consumer component and are used to predict the energy consumption values ​​of the associated consumer component.

[0016] Further important features and advantages of the invention emerge from the subclaims, from the drawings and from the associated description of the figures with reference to the drawings.

[0017] It is understood that the features mentioned above and those to be explained below can be used not only in the combination specified in each case, but also in other combinations or on their own, without departing from the scope of the present invention.

[0018] Preferred embodiments of the invention are illustrated in the drawings and are explained in more detail in the following description, wherein the same reference numerals refer to the same or similar or functionally identical components.

[0019] It shows: Fig. 1 a schematic, highly simplified and circuit diagram-like representation of a motor vehicle.

[0020] One in Fig. 1 motor vehicle 1 shown as an example has a Fig. 1. The motor vehicle 1 also has an electrical consumer component 3. The electrical consumer component 3 can be, for example, a controller 4 of the motor vehicle 1. The motor vehicle 1 also has a control device 5 which receives bus data from the bus system 2. The control device 5 predicts by means of a Fig. 1 indicated functional model 6 based on the received bus data of the bus system 2 values ​​of an energy requirement of the electrical consumer component 3. The functional model 6 is, as described below, generated and / or trained in advance using artificial intelligence.

[0021] As in Fig. 1, the functional model 6 is stored in the control device 5 in the exemplary embodiment shown. The bus data is fed to the functional model 6. The bus data can be decoded in advance using a so-called "library" stored in a decoder 11 and then made available to the functional model 6. The bus data is made available to the control device 5 permanently or at least repeatedly, so that a time-dependent bus data curve is available. The control device 5, based on the bus data curve and using the functional model 6, repeatedly predicts the associated values ​​of the energy demand of the consumer component 3. These values ​​are stored in a memory system 7 in the exemplary embodiment shown. Overall, a predicted time curve of the energy demand of the consumer component 3 is thus available.

[0022] As in Fig. 1, the values, in particular the values ​​stored in the memory system 7, can be displayed to a user (not shown) by means of a display 8. In the embodiment shown, the display 8 is one of a mobile device 9, for example a handheld device 10. As shown in Fig. 1 is also indicated, in the embodiment shown the transmission of the predicted values ​​and / or the data stored in the storage system 7 to the display 8 takes place wirelessly.

[0023] Artificial intelligence is used to pre-generate and / or train the functional model 6. A time-dependent bus data history of the bus system 2 and the associated energy consumption values ​​of the consumer component 3, determined, for example, by means of a sensor (not shown), are provided to the artificial intelligence. This means that a temporal bus data history, in particular determined in advance, and the associated energy consumption values ​​at given times are provided to the artificial intelligence as data pairs. The artificial intelligence generates and / or trains the functional model 6 using the data pairs.When generating and / or training the functional model 6, the artificial intelligence is preferably given maximum permissible inaccuracies for deviations of the predicted energy demand values ​​from the provided data pairs. The functional model 6 thus generated and / or trained is used, as described above, in the motor vehicle 1, in particular in the control device 5, to produce the energy demand values ​​of the consumer component 3 during operation of the motor vehicle 1. 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] DE 10 2021 006 280 A1

[0002]

Claims

[1] Method for predicting the energy demand of an electrical consumer component (3) in a motor vehicle (1), comprising the following procedural measures: - Generating / training a functional model (6) with an artificial intelligence such that • the artificial intelligence is provided with a time-dependent bus data history of a bus system (2) of the motor vehicle (1) to which the consumer component (3) is connected, and consumption values ​​of the energy requirement of the consumer component (3) associated with the bus data history at given times as data pairs, • the artificial intelligence generates and / or trains the functional model (6) using the data pairs, - Use of the functional model (6) in the motor vehicle (1) such that • bus data of the motor vehicle (1) are fed to the functional model (6) during operation of the motor vehicle (1), • by means of the functional model (6) from the bus data for at least two given points in time, in particular repeatedly, for the points in time associated values ​​of the energy demand of the electrical consumer component (3) are predicted, - the predicted values ​​of the energy demand and the corresponding times are stored in a storage system (7) and / or displayed to a user by means of a display (8). [2] Method according to claim 1, characterized by that when generating / training the functional model (6) a maximum permissible inaccuracy for deviations for the predicted values ​​of the energy demand from the provided data pairs is specified. [3] Motor vehicle (1) with an electrical consumer component (3) and with a bus system (2) which is connected to the consumer component (3), and with a control device (5) which receives bus data of the bus system (2) during operation, wherein a functional model (6) of the control device (5) generated and / or trained according to the method according to claim 1 or 2 is stored therein or is available to the control device (6), and wherein the control device (6) is designed such that it forecasts values ​​of the energy requirement of the electrical consumer component (3) according to the method according to one of the preceding claims by means of the functional model (6).

Images