The lack of cooperation between models and data in engineering environments reduces the innovative strength and earning capacity of the high tech sector. In addition, insufficient focus is placed on the further digitization of the complex engineering processes due to lack of insight into the relationship between these processes and their business value.
Engineering process, a complex process
In high-tech systems engineering, the relationship between the data and models of various disciplines is crucial. In current engineering practice, silos of computerization are present through a variety of tools, referred to as computer-aided design, computer-aided engineering, computer-aided manufacturing, model-driven engineering tools or language interpreters or compilers. These tools are based on generic design paradigms (e.g. solid modeling, (a-)synchronous communication) and standards used per engineering discipline to describe the design challenge at hand, while the tool technology is proprietary. The transfer of tooling results, the cross-disciplinary and cross-paradigm interaction between models and the interpretation and feedback of data generated by virtualized or operational systems are largely left to human intervention. The engineering contributions are thus loosely coupled, technology-specific, knowledge-intensive and described mostly in generic abstractions and using different paradigms.
This results in inefficient engineering processes, dominated by technology dependence, legacy issues, errors and delays in knowledge transfer between disciplines, the inability to leverage experience, unforeseen problems in operational systems. The adoption of new technologies with new design paradigms and tools is expensive and slow, requiring many engineers to adapt their skills and methods to create solutions incorporating the new technologies. The significant dependence on human contributions also makes the engineering process a complex process which is difficult to relate to business value.
Improving efficiency of engineering process
Digital engineering improves the efficiency of the engineering process through process optimization, data- and model curation, automated synthesis, design optimization and trade-offs, automated and formalized reasoning, digital twinning, sensing design and data feedback. In addition it enables the application of Artificial Intelligence (AI) in order to support the systems engineering process (AI4SE) and it provides the means and methods to allow for the development and operation of learning systems (SE4AI).
Digital Engineering implies that many of the tasks which are currently left to humans are to be computerized. Whereas humans are highly skilled and flexible when it comes to interpreting languages, a computer is not. Modeling languages are therefore very limited in terms of semantics and lack the relationship with the real world (the engineering environment). Aligning the relationship between mankind and the computer and aligning the parts of the engineering environment (disciplines and departments, but also outsourcing or purchasing building blocks) requires the correct application of context-specific and existing generic modeling languages. Languages used throughout the engineering process depend on the context of this process, the discipline or viewpoint and the level of abstraction (semantic levels).
The systems engineering community is very aware of the need to properly organize the interplay between contributions to the system. This has led to the wide adoption of system modeling languages such as SysML or its dialects and associated methods or guidance. Although these are very useful and even allow for the further integratetion of tool chains to address generic engineering problems like trade-off analysis, the lack of context awareness requires human interpretation and easily results in mismatches in the models, leading to inadequate specifications.
The application of profiles can help to better match the models to the context such that computers can be programmed to apply the right interpretations. More generally, it is important to make the use of languages and semantic levels explicit in the process to allow computerized interpretation of the models in the engineering environment. Going even further, it will require the formalization of the currently largely implicit terminology in each human-involved process step, as humans easily introduce new terminology to abstract meaining from complexity.
This coordination is an important task of future systems engineers. With the growing digitalization of engineering processes, the responsibilities of engineers will shift from performing engineering tasks to coordinating the computerized 'engineering systems' to perform the tasks. 'Systems Engineering' will increasingly transform into the orchestration of 'Engineering Systems'.