The Use of Model-Based Engineering to Design Underwater Vehicles | APPLIED RESEARCH LAB | PENN STATE UNIVERSITY
Underwater vehicles can be highly constrained, autonomous systems that are used to perform a wide variety of missions. The missions can range from long-range payload delivery to sensing missions to high-speed, rapid response missions. The vehicles are constrained by parameters such as size (diameter, length), weight, level of autonomy (self-awareness sensing and decision-making capabilities), and energy storage technologies. As a result, the vehicles are very complex and there are strong interactions between subsystems. The major subsystems include: 1) guidance, control and sensing; 2) hydrodynamics and stability; 3) energy storage and conversion; 4) propulsion; and 5) payloads.
Traditionally, system design is performed by flowing system requirements down to subsystems, enabling subsystem designers to develop point designs that satisfy these requirements. The Spiral Design process model repeats this flow-down in increasing detail in an effort to reduce development risk. It is an effective process model if the subsystems’ performance is well estimated independently of other subsystems. Unfortunately, this is seldom the case with underwater vehicles where subsystems are highly coupled.
The Applied Research Laboratory at The Pennsylvania State University has designed and built underwater vehicles for the past 70 years, most of that time following the traditional point design process. More recently, we have developed tools and methods to integrate subsystem models into a system model that enabled the definition of a broader trade space. Then teams can perform design of experiments on the system model, and analyze the experimental results. This work is motivated by the desire to aid decision-makers in understanding the impact design decisions have on complex overall system performance. Our model-based engineering process has been used on a variety of concept assessments including satellites, ground vehicles, and underwater vehicles. Our approach is to only model technologies at the level of detail needed to make current decisions, generate a broad trade space that bounds the current decision space, and make decisions to remove parts of the design space that will no longer be considered. The level of detail of the system model is then increased in subsequent design spirals to provide additional discriminating capability between the current design options and to include any validation data as it becomes available to the program.