The idea of a digital enterprise has caught traction recently as an efficient and novel means of enhancing the design, verification, validation, manufacturing, and operational processes around complex and integrated systems. The Office of Naval Research (ONR) has demonstrated interest in the digital enterprise as a means of both expanding its use of Unmanned Surface Vehicles (USVs) and scaling them to more complex use cases. The growing wave of digital engineering influence stands to augment current engineering lifecycle processes by enabling the development of digital twins and new practices around them such as virtual experimentation. These new practices will rapidly reduce the time and cost of system development and certification and, if done correctly, will accelerate the evolution of unmanned surface vehicles.

However, for these systems to be trusted to operate in the way envisioned by the ONR and larger Defense industry, several issues surrounding digital model creation must be addressed. The purpose of this study was two-fold. Firstly, to investigate a methodology for digital twin creation and virtual experimentation by developing a modelling and simulation dashboard around a prototypical unmanned surface. Secondly, to investigate potential solutions to lack of scalability and reusability in classical physics-based modelling techniques and improved digital enterprise architecting by connecting model simulation to model definition and stakeholder requirements. The first phase revolved around using both physics and data-driven information from the system to capture its behavior in three layers of interest: dynamic, electrical, and thermal. A model was created and simulated in a digital testbed to explore how improved physical and digital experimentation could reduce uncertainty in model performance. The results of this phase suggested that the spiral development approach taken to virtual experimentation platform and digital twin development could reduce the cost of system verification and validation if scaled. One part of the second phase showed that by modeling operational activities and requirements, the overall system functionality can be identified as well as any gaps in the architecture that need to be addressed. This helps identify new requirements for the USV and ensures that the process of data gathering during virtual experimentation is better understood. The structural model is then transformed into an analytical model for the actual simulation of the system. The other part of the second phase focused on causal model development using the Modelica system modelling language as a means of improving scalability. The same unmanned surface vehicle in phase one was recreated and simulated in the Dymola environment. The results were compared against experimental data from phase one and show that the Modelica model solved faster, was simpler to implement, and was more easily adapted to more complex systems than the original state-equation-based model python code.

This content is only available via PDF.
You can access this article if you purchase or spend a download.