Integrating Machine Learning in Digital Twins by utilizing SysML System Models

Wilking, Fabian; Sauer, Ch.; Schleich, Benjamin; Wartzack, Sandro

doi:10.1109/sose55472.2022.9812700

Cited by 11 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, the tool API is used to connect different tools, for example, the connection between other software, such as Cameo or Matlab Simulink (Chabibi et al 2016). Third, the system model’s XMI structure is utilized to extract the relevant information from the model (Wilking et al 2022a). The approaches are not restricted on the usage of only one class but can be assigned to multiple classes.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Packages of that model were directly derived and used as executable code to base the actual behaviour of the Digital Twin completely on the SysML model, see Table 2. In addition, a machine learning algorithm as well as relevant information for its execution was implemented into the system model to directly connect physical elements of the model with relevant elements for machine learning (Wilking et al 2022a). This enabled parallel activities to model and execute the behaviour of the Digital Twin, for example, by conducting a precise but time-consuming simulation and parallel run a machine learning script to gather less detailed but therefore quick recommendations towards the current state of the system and required actions, for example, maintenance intervals.…”

Section: System Model Utilizationmentioning

confidence: 99%

See 1 more Smart Citation

Utilization of system models in model-based systems engineering: definition, classes and research directions based on a systematic literature review

Wilking,

Horber,

Goetz

et al. 2024

Des. Sci.

Self Cite

View full text Add to dashboard Cite

The use of system models within model-based systems engineering (MBSE) is essential for improved communication or system documentation. Previous publications have investigated further reuse of these system models, for example, transforming them directly into discipline-specific models for reuse. The authors refer to this as the term “Utilization” of system models. It aims the compensation of modelling efforts and a further integration of linked models within MBSE. Motivated by a lack of common understanding of this term, a systematic literature review of the state of the art is presented. With this systematic overview, a definition and classification system for different use cases and system life cycle stages are created. These are key results to support engineers and researchers in adopting existing or discovering new utilization approaches. This supports the mission of advanced systems engineering and aims the identification of new research directions coming along with SysML v2 and the advanced systems engineering methods.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: System Model Utilizationmentioning

confidence: 99%

Utilization of system models in model-based systems engineering: definition, classes and research directions based on a systematic literature review

Wilking,

Horber,

Goetz

et al. 2024

Des. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Being utilized, these models can provide additional benefits beyond communication and documentation purposes, e.g. by providing behavioral system information for relevant software code (Wilking et al, 2022b) or defining context for the application of machine learning methods (Wilking et al, 2022a).…”

Section: Implications For An Mbse and Mbe Tool Landscapementioning

confidence: 99%

Combining System Models and CAD for Change Scenario Management

Hick,

Sanladerer,

Trautner

et al. 2023

INCOSE International Symp

Self Cite

View full text Add to dashboard Cite

In the development of complex mechatronic systems, interdisciplinary cooperation requires an exchange between stakeholders in order to ensure fulfillment of requirements and related functions. The cooperation of the stakeholders from different disciplines must be well coordinated to address systems changes. The system model as an essential aspect of model‐based systems engineering, has to be linked to specific models of involved disciplines (mechanical, electrical/electronical and software) to enable a basis for automatic identification e.g., of affected subsystems and functions. This contribution discusses the potential of model linking from the perspective of change scenarios by considering published approaches to link system models and computer‐aided design (CAD). A methodical approach on how descriptive system models can be used as base for computer‐aided design is presented. This approach is analyzed by reviewing a use case for a design change scenario of an automotive eAxle.

show abstract

“…Since the behavior usually cannot be validated with a descriptive, SysML-based system model alone [6,7], a system element needs appropriate domain models to account for all the different purposes. A domain model is, by its nature, a representation of the original system that has been shortened or abstracted in terms of scale, detail and/or functionality [8] and typically a specialized simulation model that runs in an external software tool and needs to be integrated into the SysML-based system model [6,[9][10][11][12][13]. One of the open challenges with such an integration is to provide a well-defined, modular and consistent interface between domain models and their system element counterparts, for instance to assure a correct parameter exchange.…”

Section: Introductionmentioning

confidence: 99%

Model Signatures for the Integration of Simulation Models into System Models

Zerwas

Jacobs

Kowalski

et al. 2022

Systems

View full text Add to dashboard Cite

Model-based systems engineering (MBSE) is an auspicious approach to the virtual development of cyber-physical systems. The behavior of the system’s elements is thus represented by specialized simulation models that are integrated into the descriptive SysML-based system model. Although many simulation models have been developed in research for the common system elements for various purposes and fidelities, their integration remains a major challenge: the parameter interfaces of the simulation models must be coupled with each other and with the parameters of the system elements in such a way that they are correctly parameterized. So far, this coupling can only be carried out by model experts in a time-consuming and error-prone manner. Therefore, in this paper, we first propose a concept that structures the system element parameters for targeted use in validation and design cases. Second, we propose a model signature for simulation models that differentiates its parameters by input, internal, output, and model parameters and specifies them with spatial and temporal dimensions as well as admissible ranges, among others. Based on the two contributions, domain models can be validly and automatable coupled and used for the virtual development of system elements in model-based systems engineering.

show abstract

Integrating Machine Learning in Digital Twins by utilizing SysML System Models

Cited by 11 publications

References 12 publications

Utilization of system models in model-based systems engineering: definition, classes and research directions based on a systematic literature review

Utilization of system models in model-based systems engineering: definition, classes and research directions based on a systematic literature review

Combining System Models and CAD for Change Scenario Management

Model Signatures for the Integration of Simulation Models into System Models

Contact Info

Product

Resources

About