Classification of function-oriented solution elements for MBSE

Modern products are becoming more and more integral and are often realised as cyber-physical systems (CPS). The development of CPS strongly depends on different disciplines such as mechanical, electrics/electronics and software [3, 4]. Model-based systems engineering (MBSE) is an approach to cope such a cross-domain development [5].

For developing technical systems as the drive train of an electrical vehicle, generic guidelines provide a methodological framework for a function-oriented product development process [6, 7]. However, in mechanical engineering, a function-oriented development is difficult, due to the methodical breaks in the transfer of functions into components [1]. To master these difficulties, fast functional and virtual testing provides a solution.

Virtual product testing is used to reduce development costs and time [8, 9], for instance utilising of engineering models. Those models are used to investigate the system under development concerning its physical behaviour and performance such as efficiency, loads, fatigue etc [10, 11]. Reusing already existing models helps to reduce model development time and thus, reduce development time [9].

In an existing model landscape, difficulties occur to identify a suitable engineering model out of the variety of models that exist over different solutions and different phases of the development process. Engineering models are often not structured and stored well enough to be identified for reuse.

MBSE-architectures as described in [1, 12] with the central element “solution element” contain a principle solution, domain models and workflows. By a classification scheme, domain models are classified and integrated into the solution element. This solution element provides a meaningful framework for structuring engineering models in detail. Furthermore, the solution element is the smallest and a function-oriented model unit that makes reuse during a development process possible. Function-oriented in this context means orientation towards the function of a technical system that fulfils a required customer requirement.

However, in the case of developing a drive train of an electrical vehicle, those solution elements are not identified yet. By defining solution elements, engineering models can be structured and allocated for reuse purposes. The goal of this paper is to identify mechanical solution elements as a basis to structure models systematically.

The article is structured as follows: Sect. 2 presents the state of the research of modelling a systems architecture and integrating domain models by classification schemes. In Sect. 3, the research question is defined and the solving hypothesis is stated. Section 4 describes the system of interest, which is the electro-mechanical drive train of an electrical vehicle. The approach of identifying solution elements is described in Sect. 5, whereas Sect. 6 concludes the work.

Systems Engineering is an interdisciplinary approach for developing and realising technical systems [13]. Model-based systems engineering (MBSE) is the formalised application of modelling to support system requirements, design, analysis, verification and validation within the development domain [14, 15]. A central system model is utilised and whereas the description of the system of interest is formalised. In context of MBSE, the System Modeling Language (SysML) has become established [16]. SysML is an extension of the Unified Modelling Language (UML) [17].

Several modelling methods exist to build SysML system model during product development, such as SYSMOD [18], FAS4M [19], MagicGrid [20], mecPro² [21], OOSEM [22], SPES XT [23] and motego (short cut for Model-Test-Go, formerly mse architecture) [1, 12, 24‐26]. All methods focus on the product description via requirements, functions and logic up to the product (RFLP). The approaches differ mainly in the representation of the logic layer. Only the motego method provides an adequate modelling depth for the development of contact models.

The MBSE method motego is an approach for developing cyber-physical systems with a strong focus on physical system behaviour. The approach connects the requirements level, functional level and solution level. Especially at the solution level, low fidelity up to high fidelity behaviour physical models are integrated and utilised. In Fig. 1, an example of the electrical motor within those layers is illustrated.

In the requirements layer, requirements are modelled or detailed (refine), e.g. as a text. Subsequently, the requirements are transferred into a functional architecture and are linked via satisfy connection. In the function layer, the functional architecture is defined and decomposed down to elementary functions. The elementary function is a basic function, which cannot be decomposed meaningfully and was introduced by Koller [27].

In the solution layer, specific solutions are defined concerning the previously defined functions and thus represent special cases of functions. The system solution is a specification of the functional architecture and the solution element is a specification of the elementary function. The solution element consists of the principle solution that realises an elementary function. This realisation consists of the physical effect, the active surfaces and the associated material. Furthermore, the solution element consists of domain models and workflows for model execution. The system solution is a higher-level element and contains domain models (e.g. production model or engineering models) and workflows (e.g. validation or design workflow) as well as solution elements. Engineering models are used in the development domain to execute virtual product tests such as stress tests, efficiency tests, load tests, etc. of the system under development [10, 11, 28‐30].

The different engineering models are integrated into the solution element or system solution. For integration, the engineering models are characterized by a classification scheme based on Stachowiak [32]. The classification scheme was applied to engineering models by Jacobs [1] and is characterised by the three following essential properties:

Examples of possible attributes are illustrated in Fig. 2.

In sum, the solution element is defined as the smallest element that realises on one hand a specific function (elementary function) and on the other hand integrates engineering models to test the behaviour of a specific solution of a system under development. Furthermore, the solution element offers a structuring approach to organise engineering models.

The solution element provides a function-oriented structure of engineering models for reuse and linking of engineering models with development artefacts. As a result, the reuse of solution elements shall help to increase the efficiency in product development.

However, there is no method existing to identify and define the scope of solution elements.

As seen in Sect. 2, promising approach for the mechanical domain is to structure and organise engineering models in function-oriented solution elements [1]. By that, engineering models can be identified and structured inside the element to enable efficient reuse for other solutions. However, as identified in the state of research the identification and definition of meaningful scopes for solution elements are missing.

Therefore, the research question of this paper is defined as: How can solution element be identified and defined systematically? The underlying hypothesis to answer this research question is: The interaction between two active surfaces in a contact serves as basis to identify solution elements.

Firstly, the scope of this research lies in the mechanical domain and only focuses on mechanical elements such as machine elements. Interactions within the electric motor or within the cooling system are based on magnetic fields and heat transfer and differ from mechanical interactions. Therefore, these interactions shall be investigated in a separate research.

In the current political and public discussion about the future of transport, the focus is on technical transport solutions. In passenger mobility, the focus is primarily on the e‑car and various hybrid drive systems [33]. Due to the latest relevance, the use case of an electro-mechanical drive train of an electrical urban car was chosen.

The drive train consists of different sub-systems which realise specific functions. In this paper, the focused parts of the electro-mechanical drive train are the mechanical elements. However, for the purpose of completeness, the drive unit is also introduced briefly.

The drive train is a system solution for the function provide controlled mechanical rotational energy and consist of the following subordinate system solutions:

Figure 3 shows the structure of the system hierarchically with the subordinate system solutions. The top system solution level is oriented towards the underlying technical functions of the systems, which is also a common categorisation in the literature on machine elements and drive technology (e.g. [34]). This figure applies to the specific drive train presented in this publication. However, it also applies in general to possible subordinate (mechanical) system solutions of the drive train. That means, these subordinate system solutions occur in most (mechanical) drive trains, but only differ in the characteristics or specialisations. For example, plain bearings can be used instead of roller bearings in specific applications, but this only represents a specialisation of the bearing system.

The illustrated system solutions in Fig. 3 can also be contained in other system solutions. For example, the bearing system can be integrated in the gear system as well as in the electrical machine for the shaft bearing. Here, we will not list multiple system solutions, but rather list the subordinate system solutions only once.

The aim is to identify and define solution elements systematically. The solution element is useful for reuse due to the combination of the description model (standardised description of requirements, function, etc. in a SysML-model) and the behaviour model (physical behaviour) that can be stored in SysML-model libraries. To derive solution elements, it is proposed to use the interaction of the active surfaces.

Active surfaces are surfaces of a component that participate in e.g. forces conduction. They can have different shapes and are essential for realising the system’s function. Also crucial for realisation of the function are the interactions between active surfaces. These interactions are described in physical effects that are used to achieve a certain effect.

In this paper an approach is presented, to identify and derive solution elements by a classification procedure. Using the proposed classification procedure, solution elements can be defined. It shows that only few solution elements are sufficient to describe a complex system such as the electro-mechanical drive train, as these solution elements are often repeated in a complex system. Building a library from the solution elements (e.g. a SysML-model library) is expected to amortise quickly since these few solution elements are used repeatedly. New systems are composed through different arrangements.

Engineering models can be next linked to the solution element based on an engineering model classification as stated in Sect. 2. This allows engineering models to be organised and reused in a function-oriented way.

However, a few open points remain. In the case of the identified spring, for example, the function fulfilment (save mechanical energy) does not take place in the contact, but between the two active surfaces (force transmission points) of the spring. In order to be able to map functions of this type, further description concepts are required. The concept of an “active volume” that e.g. conducts energy offers a possible solution. How well this concept can be integrated into the MBSE motego method is the subject of future research.

Furthermore, this publication only addresses mechanical contacts. In the next step, the scope will be extended to other fields such as fluidic, thermal or magnetic interactions. Here, the concept of the “active volume” is also a potential solution approach to be able to functionally describe the interactions of e.g. magnetic fields in solution elements.