Method for configuring and integrating a simulation model from a source test system into a target test system

Abstract

The invention relates to a method for configuring and integrating a simulation model (20) from a source test system (10) into a target test system (46), wherein a metadata structure (47) is created, by means of a conversion, in a neutral abstraction layer (44) from determined model configuration data and test system configuration data of a source test system (10), which metadata structure can be easily and automatically transferred, by means of a subsequent conversion, to a new target test system.

Description

[0001] DESCRIPTION

[0002] Method for configuring and integrating a simulation model from a source test system into a target test system

[0003] The invention relates to a method for configuring and integrating a simulation model from a source test system into a target test system.

[0004] Various test systems are known, particularly for testing vehicles or vehicle components, as well as their control software, control hardware, or high-voltage components. These test systems can be designed as software-in-the-loop test systems, hardware-in-the-loop test systems, physical test systems, mechanical test benches, or high-voltage test systems. All of these test systems typically include simulation models in the form of a closed control loop, via which the unit under test is connected to the rest of the system, so that the simulation and the physical test object, as well as any additional components of the test system, interact as if in a real system. For example, it is common practice in automotive test systems to combine the test object with a simulation model. An example of this would be the combination of a real engine with a simulation model for its exhaust system.

[0005] Suppliers of such test systems offer technologies for integrating various simulation models into their test systems. However, this integration requires numerous manual steps. For example, the inputs and outputs of simulation modules and their connections to the control logic and control unit of the test system must be defined via physical inputs and outputs, often with the hardware of the electronic control unit of the test object, as well as existing inputs and outputs and the connections between different models—i.e., purely virtual connections. All of these relationships are referred to as model configurations and test system configurations and must be stored in appropriate configuration software.

[0006] Such software can be divided into different layers according to the areas to be adapted. These areas include, for example, the adaptation of a simulation model's parameters—that is, specified variables that must be used as model inputs and outputs in the mathematical relationships to characterize the model being simulated—as well as the executable simulation models themselves, which can be available in various formats. Furthermore, this includes all information regarding the connections between inputs and outputs—that is, the software interfaces between simulation models and the physical test system and the device under test, as well as information regarding the existing physical connection types and their attributes. These are often implemented as bus systems or other digital or analog interfaces.

[0007] In addition, the software must have a conversion logic between the simulated physical signals and the physical inputs and outputs, whereby the signals present, for example, in a physical unit must be converted into electrical signals accordingly.

[0008] All these layers, with their interactions, parameters, and formats, form a multidimensional transformation structure.

[0009] The problem is that there is currently no standard structure for such multi-layered systems that allows simulation models to be transferred from one test system to another. Typically, the simulation parameters and the description of the model's I / O information level are defined by the test system's respective simulation tool. When the model is compiled for a target test system, all required information—i.e., all configuration and format adjustments—must be made manually to configure the target test system. This is not only very time-consuming but also prone to errors due to human intervention. Therefore, the task arises of providing a method for configuring test systems with simulation models that allows simulation models to be transferred from one test system to another as automatically as possible.

[0010] This object is achieved by a method for configuring test systems with simulation models having the features of claim 1.

[0011] The inventive method for configuring and integrating a simulation model from a source test system into a target test system comprises the following steps. First, a first model configuration and a first test system configuration of a source test system are identified in an extraction layer, and the associated first model configuration data and test system configuration data from the source test system are transferred to the extraction layer. The model configuration and the associated data are understood to mean the data relating to the simulation model, and the test system configuration data are understood to mean the data characterizing the surrounding system in which the simulation model is embedded. This can also include data relating to a test object. During the identification process, the format in which the data is available is also identified.The initial model configuration data and test system configuration data are then transferred from the extraction layer to a neutral abstraction layer. In this neutral abstraction layer, the initial model configuration data and test system configuration data are converted into metadata for automated processing. The metadata can be specified in various formats or standards. These standards ensure interoperability—that is, the exchangeability of information on the one hand, and the collaboration between two or more systems on the other. The information can therefore be used across applications. By converting data into metadata, it is possible to exchange and process it between different systems and applications without any loss of information.In this case, syntactic interoperability is particularly evident, where the collected data follows a common syntax. The XML format can serve as an example, as it defines all possible elements and the rules for linking them. This metadata is stored in a defined neutral metadata structure of the neutral abstraction layer.

[0012] In the following, a model configuration to be transferred and a second test system configuration, as well as the associated model configuration data and test system configuration data of a target test system, are identified and saved. This can be done either fully automated or manually.

[0013] By comparing the metadata of the neutral metadata structure of the source test system with the model configuration data and test system configuration data of the target test system to be configured, the model configuration data and test system configuration data to be transferred are identified. Using this data, the metadata structure of the source test system is adapted to the test system configuration and the model configuration of the target system by overwriting metadata of the source test system that is not to be transferred with metadata of the target test system to be added, thus creating a new, adapted metadata structure for the target test system. This adapted metadata structure is transferred to a transfer layer, and in the transfer layer, the metadata of the adapted metadata structure is converted into model configuration data and test system configuration data of the target test system.Finally, this converted model configuration data and test system configuration data are transferred from the transfer layer to the target test system. Using the neutral abstraction layer with its fixed metadata structure, the source test system and, if applicable, the target test system can be structurally described so that all converted data can be automatically retrieved. Additionally, processes can be initiated to implement programming instructions. All content can be evaluated, defined, and structured, and the relationships between the source test system and the target test system can be described. Accordingly, it becomes possible to automatically transfer a simulation model from a source test system to a target test system, with all usable data from the source test system being made available to the target test system without requiring any additional manual steps.This significantly reduces the configuration effort required for transferring simulation models and reliably prevents errors. Preferably, the model configuration and the test system configuration, as well as the model configuration data and test system configuration data of the target test system to be configured, are identified via the neutral abstraction layer. This eliminates the need for manual entry of this data. The entire comparison of the test systems can then be performed in the neutral abstraction layer by comparing the available metadata of the source test system and the target test system.

[0014] Advantageously, the model configuration data contains the simulated components, the inputs and outputs, and their connections of at least one simulation model. This data completely defines the simulation model. The files of the simulation model's execution software are not part of the model configuration data, but they do contain the corresponding references to these files.

[0015] In a further advantageous embodiment of the method according to the invention, the test system configuration data contains the inputs and outputs of the source test system and the target test system, as well as their connections and the existing components. This creates a complete definition of both test systems in the neutral abstraction layer. Further configuration of the target test system for integrating the simulation model is not required. The components can, in particular, include the test object, i.e., the unit under test (UUT), for example, the control unit of the test system.

[0016] To more precisely adapt the simulation model to the environment in which it is embedded, the model configuration data of a simulated component contains simulation parameters, which contain variables and simulated signals transmitted over the connections. These parameters can therefore be different in the source test system than in the target test system. Identical variables can be adopted accordingly, while others must be adapted to the target test system.

[0017] In a further embodiment, the simulation parameters contain the mathematical relationships stored in the simulation model in the form of equations or maps, as well as the basic values ​​to be defined and the values ​​determined in the model. Thus, a complete representation of the simulation model is stored in a clearly defined form.

[0018] Furthermore, it is advantageous if the test system configuration data contains the component parameters, which contain the variable base values ​​and the signals transmitted via the connections of the determined values ​​of the original test system. This is necessary to enable a complete comparison of the two test systems and to ensure that parameters and signals present in both test systems can be transferred automatically.

[0019] Preferably, the model configuration data and test system configuration data are read out using a parsing algorithm and made available for transmission. Such a parsing algorithm is used to decompose and convert the configuration data into a different format—in this case, to decompose and convert the metadata format. Conversely, XML parsers can be used to analyze the metadata and make the information it contains available to the target test system.

[0020] In a further advantageous embodiment, the metadata structure of the source test system and / or the second model configuration and the adapted metadata structure of the model configuration data to be configured and the test system configuration data of the target test system are stored in the neutral abstraction layer. This makes it possible to use the acquired and additionally inserted data when a model is transferred from the source test system to the target test system again, thereby further reducing the transfer time. The metadata structure of the source test system can also be used for transfer to other target test systems, or the second model configuration and the second test system configuration with their model configuration data and test system configuration data to be configured can be used for transferring simulation models of other source test systems to the same target test system.In particular, the connections between the model configuration data and the test system configuration data preferably also include connections between different simulation models, as well as between the target test system and the source test system and the simulation models. In this way, different simulation models of different source test systems can also be linked to one target test system.

[0021] The method according to the invention enables, for the first time, the transfer of simulation models from one test system to another in a largely automated manner, independent of the manufacturer of the test systems or the simulation models. By abstracting all hardware-, test system-, and test system-software-specific configuration information, this data is made available independently in a readable, standardized form, allowing the user to easily transfer system configurations from one test system to another, with all format and configuration adjustments being made automatically.

[0022] The procedure is described below using the example of transferring a simulation model from an electric motor test bench to an inverter test bench.

[0023] The figure shows a sketch of the process according to the invention.

[0024] The figure shows an initial test system 10 in the form of a test bench for testing an electric motor as test object 12. This consists of a physical part 13 of the test bench with a first test system configuration 14, which has a dynamometer 16 for measuring the load of the test object 12. Furthermore, the initial test system 10 consists of a first model configuration 18 of a simulation model 20, via which a vehicle 24, a driver 26 of the vehicle 24, a virtual control unit 28, a battery 30, and a drive train 32 are simulated as simulated components 22. The test object 12 and the dynamometer 16 are controlled via the simulation model 20, and the data of the test object 12 is processed. Such a simulation model 20 can be available, for example, as a functional mock-up, in Simulink file format, or in ecp file format.The output test system 10 has various inputs and outputs 34, 35, 36, 37, namely, on the one hand, inputs and outputs 34, 35, 36 between the simulation model 20 and the physical part 13 of the test bench, and, on the other hand, virtual inputs and outputs 37 within the simulation model 20 between the simulated components 22. In addition, this output test system 10 has corresponding connections 38 between the inputs and outputs 34, 35, 36, 37. These thus include the software wiring information between the simulation models 20, the simulation model 20 and the physical part 13 of the test bench, including the test object 12, from the physical part 13 of the test bench, including the test object 12, to the simulation model 20, but also the physical connections from the simulation model 20 to the physical inputs and outputs 34, 36.Typically, communication bus inputs and outputs such as CAN and CAN parameters, EtherCAT and EtherCAT parameters, or analog or digital general-purpose inputs or outputs or pulse width modulation generators are used here.

[0025] The model configuration 18 contains various simulation parameters 40. These simulation parameters 40 include, on the one hand, variables of the respective simulated component and, on the other hand, their transmitted signals. The simulation parameters 40 thus include the mathematical relationships stored in the simulation model 20 in the form of equations or maps, such as the speed-torque map of the electric motor stored in the control unit 28, the pedal map, or the coast-down test equation, as well as the basic values ​​to be defined, such as the vehicle mass, the shaft stiffness, the damping, and the mass inertia of the simulated drive train 32. These simulation parameters 40 can thus be adapted from application to application with values ​​desired by the user.The simulated signals contain, for example, predefined signals regarding existing temperatures, speeds or similar, which are either present as a predefined value or are determined in the simulation.

[0026] The test system configuration data also contain the parameters of the components, in this case the dynamometer or the electric motor to be tested, which contain their variable basic values ​​and the signals of the determined values ​​of the initial test system that may be transmitted between them via the connections.

[0027] All of the aforementioned signals are transmitted via the aforementioned inputs and outputs 34, 35, 36, 37 and the associated connections 38 within the output test system 10, with the respective signals being assigned to the corresponding connections 38. Thus, the simulated drivetrain component 32 has an output 34, which is connected via a connection 38 to an input 37 of a controller 41 of the dynamometer 16. The transmitted variable is the setpoint speed. Furthermore, the simulated drivetrain component 32 has an input 35, via which the resulting torque is transmitted from the test object 12. The simulated component of the virtual control unit 28 has an output 36, via which the test object is supplied with a setpoint for the drive torque.

[0028] All these relationships and values ​​thus form the model configuration data and test system configuration data of the model configuration 18 and the test system configuration 14, respectively, and are stored in a configuration software.

[0029] The model configuration 18 and the test system configuration 14 are first identified in an extraction layer 42 and the corresponding determined model configuration data and test system configuration data are collected and stored in the present form.

[0030] The model configuration data and test system configuration data are then transferred to a neutral abstraction layer 44, which may also contain the extraction layer 42. In this layer, all this data is abstracted by converting it into neutral and generally defined metadata that can be expanded and adapted for a target test system 46. The resulting metadata structure 47, with its standardized format, decouples all format, configuration, and workflow dependencies from the test systems from which the data originate. Accordingly, it enables the configuration to be performed in a neutrally designed configuration tool. In the present case, a comparison is thus made between the model configuration data and test system configuration data of the source test system 10 and the target test system 46, which in this case concerns a test bench for an inverter, forming the test object 48.This target test system 46 again consists of a physical part 50 of the test bench with a first test system configuration 52, which has a dynamometer 54 and an electric motor 56. Furthermore, the target test system 46 is to be equipped with a desired model configuration 58 of the simulation model 20, which is to be configured, via which the vehicle 24, a driver 26 of the vehicle 24, a virtual control unit 28, a battery 30, and a drive train 32 are to be simulated as simulated components 60. However, when testing the inverter as the test object 48, this drive train 32 must be divided into a transmission model 61 and a drive train model 63 forming the remaining drive train. Furthermore, this model configuration 58 must also contain various simulation parameters 62.While most of these simulation parameters 62 and the existing inputs and outputs 64, 65, 66, 67, and connections 68 correspond to the simulation parameters 40 and the inputs and outputs 34, 36, and connections 38 of the source test system 10, others must be replaced. The model configuration 58 and the test system configuration 52 of the target test system 46 must be identified accordingly. This can be done manually or, at least partially, also by using the neutral abstraction layer 44 by reading in the existing test system configuration data of the target test system 46 and converting it into a metadata structure 47.

[0031] The identification of these data to be replaced, deleted or supplemented is carried out by comparing the first model configuration 18, the simulation parameters 40 and the existing inputs and outputs 34, 35, 36, 37 as well as the connections 38 of the initial test system 10 with the desired model configuration 58, simulation parameters 62 as well as the inputs and outputs 64, 65, 66, 67 and the connections 68 of the target test system 46.

[0032] In the present case, in addition to the aforementioned replacement of the drivetrain model 32 with the new drivetrain model 63 and the transmission model 61, the variables to be transmitted with the associated inputs and outputs 64, 65, 66, 67 must also be changed, since the source test system 10 is torque-controlled, while the target test system 46 is speed-controlled. Thus, the output 64 of the virtual control unit 28 must now be connected to the test object 48, i.e., the inverter, in order to transmit the torque setpoint. The output 34, which was connected via a connection 38 to the input 37 of the controller 41 of the dynamometer 16, must now be replaced by an output 65 of the transmission model 61. The variable to be transmitted via this output must also be replaced, because the torque must now be transmitted via this output 65 instead of the dynamometer speed.The input 35 of the simulated drive train component 32 of the output test system 10 must also be replaced, since this input 66 must now be assigned to the transmission model 61 and a speed of the electric motor is now reported back instead of the engine torque.

[0033] For the simulation parameters 40, the mathematical relationships remain in the form of equations or maps. However, it is necessary to change some simulation parameters 70 for the target test system 46, namely the basic values ​​for shaft stiffness and damping, and to delete the mass inertia.

[0034] These simulation parameters 70 are thus adapted to the values ​​desired by the user in the metadata structure 47 of the neutral abstraction layer 44. This then contains a generally valid, adapted metadata structure 71 for the target test system 46. These now existing model configuration data and test system configuration data can now be moved as metadata to a transfer layer 72, where they are converted into the format required in the target test system.

[0035] In this way, the simulation model 20 of the source test system 10 is transferred completely and adapted to the target test system 46. No further manual intervention is required. The metadata structure 47 of such a moved simulation model 20 is saved and can thus also be used for transfer to other test systems. The corresponding metadata structure 47 of the target test system 46 can also be stored in the neutral abstraction layer 44. In the same way, the simulation models of any test systems can of course be moved, even between software-in-the-loop test systems, hardware-in-the-loop test systems at the signal level, and physical test systems such as mechanical test systems or high-voltage test systems. In this case, virtual inputs and outputs may also be replaced by real ones if necessary. This is done using the metadata structure, in which no distinction is necessary in this regard.

Claims

PATENT CLAIMS 1. A method for configuring and integrating a simulation model (20) from a source test system (10) into a target test system (46), comprising the steps: Identifying a first test system configuration (14) and a first model configuration (18) of an initial test system (10) in an extraction layer (42), Transferring initial model configuration data and Test system configuration data from the initial test system (10) into the extraction layer (42), Transferring the first model configuration data and test system configuration data from the extraction layer (42) to a neutral abstraction layer (44) in which the first model configuration data and test system configuration data are converted into metadata, Storing the metadata in a defined neutral metadata structure (47) of the neutral abstraction layer (44), Identifying and storing a model configuration (58) to be transferred and a second test system configuration (52) and model configuration data to be configured and test system configuration data of a target test system (46), Identifying model configuration data and test system configuration data to be transferred by comparing the metadata of the neutral metadata structure (47) of the source test system (10) with the model configuration data and test system configuration data to be configured of the target test system (46), Adapting the metadata structure (47) to the test system configuration (14) and the model configuration of the target test system (46) by overwriting metadata of the source test system (10) that is not to be transferred with the metadata of the target test system (46) to be added, Transferring the adapted metadata structure (71) of the target test system (46) into a transmission layer (72), Converting the transmitted metadata of the target test system (46) in the transmission layer (72) into model configuration data and test system configuration data of the target test system (46), Transferring the converted model configuration data and test system configuration data from the transmission layer (72) to the target test system (46).

2. A method for configuring and integrating a simulation model from a source test system into a target test system according to claim 1, characterized in that the identification of the model configuration (58) and the test system configuration (52) and the model configuration data and test system configuration data of the target test system (46) to be configured takes place via the neutral abstraction layer (44).

3. Method for configuring and integrating a simulation model from a source test system into a target test system according to claim 1 or 2, characterized in that the model configuration data contain the simulated components (22, 60), the inputs and outputs (34, 35, 36, 37, 64, 65, 66, 67) and their connections (38, 68) of the at least one simulation model (20).

4. Method for configuring and integrating a simulation model from a source test system into a target test system according to one of the preceding claims, characterized in that the test system configuration data contain the inputs and outputs (34, 35, 36, 64, 65, 66) of the source test system (10) and the target test system (46) as well as their connections (38, 68), and the existing components (22, 60).

5. A method for configuring and integrating a simulation model from a source test system into a target test system according to one of the preceding claims, characterized in that the model configuration data of a simulated component (22, 60) comprise simulation parameters (62) which contain variables and simulated signals transmitted via the connections (38, 68).

6. Method for configuring and integrating a simulation model from a source test system into a target test system according to claim 5, characterized in that the simulation parameters (62) contain the mathematical relationships stored in the simulation model (20) in the form of equations or maps as well as the basic values ​​to be defined and the values ​​determined in the simulation model (20).

7. Method for configuring and integrating a simulation model from a source test system into a target test system according to one of the preceding claims, characterized in that the test system configuration data comprise the parameters of the components which contain the variable basic values ​​and the signals of the determined values ​​of the source test system (10) transmitted via the connections.

8. A method for configuring and integrating a simulation model from a source test system into a target test system according to one of the preceding claims, characterized in that the model configuration data and test system configuration data are read out via a parsing algorithm and made available for transmission.

9. Method for configuring and integrating a simulation model from a source test system into a target test system according to one of the preceding claims, characterized in that in the neutral abstraction layer (44) the metadata structure (47) of the source test system and / or the adapted metadata structure (71) of the configuring model configuration data and test system configuration data of the target test system (46) are stored.

10. A method for configuring and integrating a simulation model from a source test system into a target test system according to one of the preceding claims, characterized in that the connections (38, 68) of the model configuration data and Test system configuration data also include connections (38, 68) between different simulation models (20) as well as between the target test system (46) and the source test system (10) and the simulation models (20).

Images