Tradespace Exploration of a UAV Conceptual Design Using Model-Based Systems Engineering

The objective of this research is to apply model-based systems engineering approaches to the conceptual design of unmanned aerial vehicles. This is accomplished by evaluating the models of aircraft performance, extracting input and output parameters from the models, creating chains of models, and implementing the models in the MATLAB programming language. By following this process, it is possible to identify the global parameters that remain constant across models, the shared input parameters, the dependencies between models, and the feedback loops with the systems models. The models are currently implemented in two files using Microsoft MS Excel, one is focused on constraint sizing and the other file is focused on weight and energy sizing. The worksheets and associated models are analyzed and several shortcomings are identified. First, there were several models that shared parameters which were often repeated and not linked within the implementation. In the MS Excel tool, there are 63 models and 83 parameters with 21 user entered parameters, out of which 13 parameters were shared in between two worksheets. These parameters and models can lead to inconsistent values which propagate across the models. Second, there were several additional parameters that were identified which did not contribute to the final calculation. The number of models were reduced from 63 to 43 models and parameters from 83 to 51 with the same 21 user entered parameters. The models were implemented in MATLAB and verified through four UAV conceptual design exercises. The outcomes from this research provide a basis for formalizing models commonly used in conceptual aircraft. This will lead to better structured framework for the models to be executed in and will reduce inconsistency and errors. Specifically, the model-based approach provides a simplified method to calculate the design parameters. The connectivity amongst the models helps the designer to connect various repeating models and to help avoid calculation error. The modular nature of the models helps the designer to interchange models based on the mission statement requirements. This leads to reduced number of variables and increased consistency of the output parameters. Finally, the model-based approach enables tradespace exploration to be more easily completed as compared to the MS Excel-based tool