An object-oriented analysis and optimization control environment for the conceptual design of aircraft

TR Number
Journal Title
Journal ISSN
Volume Title
Virginia Tech

This dissertation presents a new software environment for performing design activities utilizing numerical optimization algorithms. Specifically, the research focuses on new techniques for interacting with optimization software, unique methods for handling design information, and creative methods for visualizing optimization results.

The application of computer aided optimization algorithms in the engineering design process is hindered by a lack of software methodology and subsequent tools that give the designer adequate control of the process. This process includes setting up the problem, executing the problem with proper key feedback, and extracting the pertinent information from the results to make sound engineering decisions. Elements of data management and visualization are critical to supporting this decision making process.

Numerical optimization can be a powerful tool for the design engineer. If applied properly, vast savings in time and analysis effort can be realized. All too often, however, optimization is under-utilized because of lack of trust for the methods, or, more commonly, a lack of understanding for why the design arrived at the final result. The optimization approach and supporting software tools developed in this research provide a system that offers the design engineer insight through visualization into the complete history of the optimization problem.

For the designer to understand the results from the optimization process, he must be presented with a traceable path of changes to the design and what the influences were at each change. Considerable time is spent disseminating the results of a problem after the optimization algorithms are utilized. Results are scrutinized and histories are examined in great detail to understand why the process arrived at the resulting design.

In addition to presenting comprehensible results of the optimization routines, sophisticated software control of the optimization of complex engineering systems is necessary. The algorithms as well as the actual problem formulation must be accessible and controllable for the design engineer to fully realize the capability of numerical optimization.

The base treatise proposed herein is divided into three distinct areas:

• Software tools for improving process interaction and feedback • A mathematical strategy for gaining insight from process information • Illustrative examples of how the developed methods are employed