Search Results

This paper critically evaluates the use of Neural Networks (NNs) as metamodels for design applications. The specifics of implementing a NN approach are researched and discussed, including the type and architecture appropriate for design-related tasks, the processes of collecting training and validation data, and training the network, resulting in a sound process, which is described. This approach is then contrasted to the Response Surface Methodology (RSM). As illustrative problems, two equations to be approximated and a real-world problem from a Stability and Controls scenario, where it is desirable to predict the static longitudinal stability for a High Speed Civil Transport (HSCT) at takeoff, are presented. This research examines Response Surface Equations (RSEs) as Taylor series approximations, and explains their high performance as a proven approach to approximate functions that are known to be quadratic or near quadratic in nature.

A technique is proposed for examining complex behaviors in the “pilot - vehicle - operational conditions” system using an autonomous situational model of flight. The goal is to identify potentially critical flight situations in the system behavior early in the design process. An exhaustive set of flight scenarios can be constructed and modeled on a computer by the designer in accordance with test certification requirements or other inputs. Distinguishing features of the technique include the autonomy of experimentation (the pilot and a flight simulator are not involved) and easy planning and quick modeling of complex multi-factor flight cases. An example of mapping airworthiness requirements into formal scenarios is presented. Simulation results for various flight situations and aircraft types are also demonstrated.

The ultimate goal of knowledge-based aircraft design, pilot training and flight operations is to make flight safety an inherent, built-in feature of the flight vehicle, such as its aerodynamics, strength, economics and comfort are. Individual flight accidents and incidents may vary in terms of quantitative characteristics, circumstances, and other external details. However, their cause-and-effect patterns often reveal invariant structure or essential causal chains which may re-occur in the future for the same or other vehicle types. The identification of invariant logical patterns from flight accident reports, time-histories and other data sources is very important for enhancing flight safety at the level of the ‘pilot - vehicle -operational conditions’ system. The objective of this research project was to develop and assess a method for ‘mining’ knowledge of typical cause-and-effect patterns from flight accidents and incidents.