Validating model-based design simulation: The impact of abstraction and fidelity levels

For many complex engineered systems, a risk informed approach to design is critical to ensure both robust safety and system reliability. Model-based design enables this through early design stage failure analysis. This paper describes initial work in the use of an experimental testbed to validate a function-based failure analysis method. The behavioral abstraction and model fidelity levels that allow this method to be used in the early design stages ultimately affect the validation metrics needed to understand and quantify the uncertainty of that analysis. To that end, in this paper, we present the computational and experimental simulation of failure scenarios on a robotic arm. This analysis provides insights into how behavioral abstraction relates to the type of behaviors that can be validated and how different model fidelity levels can be related to each other. Ultimately, the goal of this work is to develop a method and set of metrics for validating design-stage, model-based failure analysis. This research effort will enable objective comparison and certification of these design methods.