Gregor Höpfner

Future cyber-physical systems (CPSs), integrating subsystems of the mechanical, electrical and software domains, are becoming increasingly interconnected and complex. As complexity grows, testing effort increases as well. This includes the test-case definition step, where the test targets and boundary conditions are specified. With rising system complexity, the effort required to ensure that all relevant conditions for each test target are identified increases. Manual test-case definition remains the norm, creating effort bottlenecks in ensuring systematic coverage and compliance with standards such as ISO 26262 and ISO 29119. This paper explores how large language models (LLMs) can support the identification of complex boundary conditions for CPS test cases through detailed requirement analysis. The impact of performing taxonomy-guided, structured requirement mapping prior to test-case generation was evaluated by comparing it with a version without this guidance. Furthermore, the influence of supplying a Model-Based Systems Engineering (MBSE) system model as context information via Graph RAG is examined. The results show that structured, stepwise reasoning significantly improves reliability and consistency over unguided generation, while system-model information provides valuable contextual insight but has a minor impact in the chosen example. These findings outline a scalable framework for AI-assisted test-case generation.

The complexity of modern cyber–physical systems is steadily increasing as their functional scope expands and as regulations become more demanding. To cope with this complexity, organizations are adopting methodologies such as model-based systems engineering (MBSE). By creating system models, MBSE promises significant advantages such as improved traceability, consistency, and collaboration. On the other hand, the adoption of MBSE faces challenges in both the introduction and the operational use. In the introduction phase, challenges include high initial effort and steep learning curves. In the operational use phase, challenges arise from the difficulty of retrieving and reusing information stored in system models. Research on the support of MBSE through artificial intelligence (AI), especially generative AI, has so far focused mainly on easing the introduction phase, for example by using large language models (LLMs) to assist in creating system models. However, generative AI could also support the operational use phase by helping stakeholders access the information embedded in existing system models. This study introduces an LLM-based multi-agent system that applies a Graph Retrieval-Augmented Generation (GraphRAG) strategy to access and utilize information stored in MBSE system models. The system’s capabilities are demonstrated through a chatbot that answers questions about the underlying system model. This solution reduces the complexity and effort involved in retrieving system model information and improves accessibility for stakeholders who lack advanced knowledge in MBSE methodologies. The chatbot was evaluated using the architecture of a battery electric vehicle as a reference model and a set of 100 curated questions and answers. When tested across four large language models, the best-performing model achieved an accuracy of 93 percent in providing correct answers.

The complexity of modern cyber-physical systems is steadily increasing as their functional scope expands and as regulations become more demanding. To cope with this complexity, organizations are adopting methodologies such as Model-based Systems Engineering (MBSE). By creating system models MBSE promises significant advantages such as improved traceability, consistency, and collaboration. On the other hand, the adoption of MBSE faces challenges in both the introduction and the operational use. In the introduction phase, challenges include high initial effort and steep learning curves. In the operational use phase, challenges arise from the difficulty of retrieving and reusing information stored in system models. Research on the support of MBSE through Artificial Intelligence (AI), especially Generative AI, has so far focused mainly on easing the introduction phase, for example by using Large Language Models (LLM) to assist in creating system models. However, Generative AI could also support the operational use phase by helping stakeholders access the information embedded in existing system models. This study introduces an LLM-based multi-agent system that applies a Graph-Retrieval-Augmented-Generation (GraphRAG) strategy to access and utilize information stored in MBSE system models. The system’s capabilities are demonstrated through a chatbot that answers questions about the underlying system model. This solution reduces the complexity and effort involved in retrieving system model information and improves accessibility for stakeholders who lack advanced knowledge in MBSE methodologies. The chatbot was evaluated using the architecture of a battery electric vehicle as a reference model and a set of 100 curated questions and answers. When tested across four large language models, the best-performing model achieved an accuracy of 93 percent in providing correct answers.

Rising consumer demand and stricter regulations regarding sustainability are leading to an increasing consideration of sustainability aspects during product development. To evaluate environmental impacts, life cycle assessments (LCAs) are established as the most comprehensive method. However, the extensive data-need for setting up the life cycle inventory (LCI) model limits their widespread application. Moreover, LCAs are normally conducted post-development, offering no support in decision making during the development process. An approach is needed that provides product-specific information during product development to be used in LCAs to allow the execution of LCAs already during product development. Model-based systems engineering (MBSE) promises to meet this need by providing a framework for seamless interdisciplinary product development. A central element in many MBSE approaches is a SysML system model that describes the architecture and behavior of the product and provides all relevant information for needed domain-specific models. In this contribution, a novel approach is presented utilizing a SysML system model to automatically execute an impact assessment within an LCA tool. For this, the minimum data-need for setting up LCI models in established LCA tools is analyzed, and a framework for modeling life cycles in SysML system models is presented. To allow an automated execution of an LCA, an interface between a SysML tool and an LCA tool is implemented. In the example of a laser-structured metal part, this contribution shows that SysML system models are suited to provide the needed product-specific data for setting up LCI models and that they can be utilized to execute LCAs already during product development.

Today’s systems are becoming increasingly complex due to the multitude of interactions between subsystems. This is also true for the electromechanical drivetrain and its physically interacting cooling system. In order to provide a virtual representation of such systems, including system architecture and product behaviour, model-based systems engineering (MBSE) introduces system models. System models are built using system elements and reoccurring models. MBSE, therefore, enhances the efficient development of complex systems by promoting model reuse in interdisciplinary architectural modelling. The reuse of models, such as calculation models, reduces redundancy, accelerates development iterations, and streamlines consistency. However, there is a lack of standardised and reusable model libraries to facilitate this reuse. In the approach in this paper, the reusability of those models is facilitated by the system elements, referred to as “solution elements”. MBSE system elements enable the structuring, reuse, and organization of models within model libraries. The identification of these system elements for heat-exchanging systems, however, remains an open challenge. Consequently, the aim of this paper is to develop a method for systematically identifying system elements in heat-exchanging systems, providing a formalized approach to reusing thermal models. The method focuses on functional and heat-transfer processes at the contact level referred to here as thermal contacts. The developed method is demonstrated through a case study of a thermal management system (TMS) of an electric truck. It is shown that a small set of recurring system elements can be used to represent a large number of individual thermal interactions, within TMS components and, therefore, streamline modelling efficiency significantly.

One approach to tackle the complexity in the development of cyber-physical systems is system decomposition and modelling in a central system model with model-based system engineering (MBSE). However, MBSE also requires increased development effort due to the high degree of formalisation of the development results using the Systems Modelling Language (SysML). One solution is reusing system elements from model libraries, which requires formalization that allows them to be networked into consistent models. Due to the abstract nature of SysML, the options for formalizing system elements are extremely diverse, meaning that even identical technical system elements are modelled with a different structure by engineers. As a consequence, no standardized structure is maintained and the reuse of system elements from model libraries is very unlikely. This leads to high, redundant modelling efforts as well as a considerable error potential and significantly hinders the economic application of MBSE in mechanical engineering. A standardized structure of the system elements can be ensured through MBSE methodology specific modelling guidelines. Current modelling guidelines for MBSE are based on the SysMLv1 language standard and cannot be used unchanged in SysML v2. Nevertheless, SysML v2 offers great potential for modelling reusable system elements, especially with its synonymous textual and graphical modelling capabilities, enabling the easy linkage of knowledge from various domain-specific tools. Consequently, modelling guidelines based on SysML v2, which guarantee a standardized structure, are required to fully exploit the potential of model libraries in MBSE. To solve this deficit, this paper presents modelling guidelines based on SysML v2 and their application for building reusable system elements.

As product development cycles become increasingly short while products grow more complex, new approaches in product development are needed to accommodate these developments. One of these approaches is Model-based Systems Engineering (MBSE). MBSE enables the representation of products, which are seen as systems, in one seamless system model. To do so the product and its different aspects, such as physical behavior, are described by different models, that are then interconnected with each other. One advantage of applying MBSE is, that the usage of physical behavior models allows efficient virtual validation of many different product features and evaluation of implication due to design changes. However, performing sufficiently accurate virtual tests, especially for mechanical properties, requires detailed behavior models, which leads to a high modeling effort. For this reason, in the past efforts were made to standardize model components, referred to as system elements, that could be reused and thus reducing the modeling effort. The aim is to enable the modeling of any technical product from the composition of these base components. However, these system elements have not yet been modeled and organized in a library, which could then be distributed across modelers. For this reason, the aim of this paper is to firstly create a model library based on previously defined system elements that satisfies the required physical model properties. Furthermore, the applicability of this library in modeling is investigated and improvement potentials are derived. To do so a simplified gearbox model is created from multiple elements contained in a predefined library. The model created is then used to specifically investigate the ability to verify static safety using the physical behavior models contained in the library. The result of this work is an expandable model library of reusable system elements, that allows the aggregation of system models, representing products and the virtual verification of these products. Thus, enabling for a faster product development process as the modeling effort for the application of MBSE is reduced.

New emission regulations and the associated mobility transition has led to an increased demand for new drivetrain concepts. These new cross-domain concepts, such as electromechanical drivetrains, are characterized by a high number of interactions between components which must be tested to ensure the functionality of the developed system. To achieve a short time-to-market, it is necessary to test subsystems and components early in the development process. Although virtual testing through simulation offers cost-effective solutions, physical testing remains essential for validating real physical behavior. Nevertheless, due to high costs and disparate delivery times of components, physical testing alone is insufficient. To counteract, Hardware-in-the-Loop (HiL) testing can be performed to combine the benefits of physical and virtual testing in the early stages of development, particularly when not all components are yet physically available. However, the problem arises, that for HiL testing a virtual HiL model must be available which suits the individual test case consisting of the Device Under Test (DUT) and the associated purpose of the test including the test bench on which the test will be performed. These models must be created with high manual effort. To counteract reusability can be achieved by using predefined model elements. Therefore, we present an approach for creating configurable HiL models. For this approach, a vehicle model consisting of standardized, modular model element for each subsystem is set up to be used for HiL testing. Modularization of models ensures quick and easy adjustment and efficient product development which is also beneficial for the creation of HiL models. The created vehicle model serves as a basis for all required HiL models which can then configured. For the configuration a workflow is developed which allows an automatic derivation of the required HiL model by considering the defined test case.

Modern vehicles are complex cyber-physical systems with numerous customer functions and operating states, requiring careful management during development. Ensuring that all potential states and transitions are rigorously defined at the requirement stage is critical for vehicle safety and reliability. Due to stringent regulatory requirements and competitive market pressures, it is essential to articulate these requirements unambiguously, free of redundancy, and correctly. However, even a formally correct set of requirements can be underspecified, leading to unintended operating states that compromise vehicle functionality. To ensure requirement completeness, various approaches have been proposed to generate operating states and transitions allowed by a given set of requirements, ensuring that only intended behaviors are permitted. Existing methodologies often rely on natural language requirements, prone to ambiguities and errors. Furthermore, several methods formalize requirements using temporal logic languages, which can be checked algorithmically. Because of the complexity of modern vehicles, Model Based Systems Engineering (MBSE) has become state of the art in vehicle requirements engineering, as MBSE enables rigorous traceability of requirements and early model-based validation of requirement fulfillment. Currently a gap remains in integrating the comprehensive generation of operating states and transitions with the formalization in temporal logic languages, specifically for the use in MBSE development environments. This paper introduces a multi-step approach that formalizes requirements using Linear Temporal Logic (LTL) and generates operating states and transitions, which are then modeled in a SysML v2 state machine diagram, enhancing both the development process and the system’s reliability.