Top-down development following the V-model of systems engineering can help to deal effectively with uncertainty in systems design without a particular uncertainty model. Often in industrial practice, however, concrete design steps are difficult to identify using the general theory – and the V-model remains a theoretical construct. This paper presents a simple but effective framework for systems engineers to connect the V-model theory with quantitative design methods, thus enabling a structured process for the systematic distributed design of large multi-disciplinary systems subject to uncertainty. The framework proposed includes three distinct steps: first, the system structure is modelled by specifying all relevant dependencies between system variables (i.e. design variables and objective quantities) in a hierarchical dependency graph. This simple formalism provides a concrete structure for the design process. Second, quantitative bottom-up mappings between system variables are established by physical or mathematical models. Third, quantitative top-down mappings are used to provide regions of permissible designs, so-called solution spaces. They encompass variability related to epistemic uncertainty and are maximised for the integration of requirements from different disciplines. They are computed by existing numerical algorithms or a projection technique. Several vehicle design problems demonstrate the general applicability and the effectiveness of the approach.

Industry 4.0, cyber-physical systems, and digital twins are generating ever more data. This opens new opportunities for companies, as they can monitor development and production processes, improve their products, and offer additional services. However, companies are often overwhelmed by Big Data, as they cannot handle its volume, velocity, and variety. Additionally, they mostly do not follow a strategy in the collection and usage of data, which leads to unexploited business potentials. This paper presents the implementation of a Digital Twin module in an industrial case study, applying a concept for guiding companies on their way from data to value. A standardized use case template and a procedure model support the companies in (1) formulating a value proposition, (2) analyzing the current process, and (3) conceptualizing a target process. The presented use case entails an anomaly detection algorithm based on Gaussian processes to detect defective products in real-time for the extrusion process of aluminum profiles. The module was initially tested in a relevant environment; however, full implementation is still missing. Therefore, technology readiness level 6 (TRL6) was reached. Furthermore, the effect of the target process on production efficiency is evaluated, leading to significant cost reduction, energy savings, and quality improvements.

Design for optimal commonality in product families is different from design for optimal performance. While optimal performance may be achieved by the choice of appropriate design parameter values for all system components, optimal commonality requires a particular scheme of sharing components among systems. The number of possibilities to share components can be quantified by Bell's number and becomes large quickly, thus making optimization extremely expensive. This paper presents an approach to identify components that may be shared in order to optimize commonality for a product family of arbitrary high-dimensional nonlinear systems. Solution spaces are computed for each system using iterative Monte Carlo sampling. On these solution spaces, all design goals are reached. They are expressed as sets of permissible intervals for all design parameters. When parameter intervals from different systems overlap, they may assume the same value and components may be shared. The approach is applied to vehicle chassis design. A set of common components is computed for 13 vehicles with ten design parameters each, such that all design goals are satisfied and the number of different component designs is small.