Elements of computational flight dynamics for complete aircraft

Abstract

This work summarizes the effort to introduce high–fidelity methods in aircraft design process. Flight dynamic characteristics are crucial in aircraft design. Simulation tools evaluate the aircraft statical and dynamical stability and manoeuvrability usually are based on a previously computed tabular aerodynamic model.

During the conceptual design phase, the aerodynamic database is usually computed with semi–empirical methods. These tools rely on existing configurations databases (statistical methods) or linear aerodynamic hypothesis (e.g. vortex lattice method), and so are not suitable for innovative designs. The exploitation of such methods may lead to evaluation errors in the design process, which can be found only in the following steps and so may be very expensive to rectify via additional work, wind tunnel and flight testing, enlarging the time–to–market and increasing the whole life cycle product cost.

The adoption of high–fidelity, physical based aerodynamic models starting from the very first steps of the aircraft design would reduce the uncertainty of current design procedures and prevent costly aircraft retrofitting. Computational fluid dynamics may be utilized to achieve the required high–fidelity, but, because of the substantial computational cost, it is currently used only during ensuing design steps. In this thesis the steps towards an autonomous high–fidelity flight dynamics analysis are presented.

A tool for generating the aerodynamic tables with the semi–empirical United States air force stability and control digital compendium with the common parametric aircraft configuration schema is developed. The function for the flow solver Edge is updated and both scripts are implemented and validated inside the computerised environment for aircraft synthesis and integrated optimisation methods.

Reduced order models to overcome computational fluid dynamics limitations for automated generation of aerodynamic tables are then presented. Two methods are developed in order to obtain a more efficient approach for samples positioning inside the flight envelope domain. Emphasis is given on the ability to capture nonlinearities appearance in the flow field with only a few computations over the whole flight envelope. The methods rely on Kriging interpolation, and are validated for semi–analytical functions and for real test cases. This may permit to reduce the number of required computational fluid dynamics solutions to use the flight simulator of a factor of some tens, without compromising the main aircraft statical and dynamical behaviour results.

A test case is then presented, showing the statical and dynamical aircraft stability comparison between different geometry configurations, by use of reduced order models and with a low computational budget. The limitations of a derivatives based aerodynamic model are then presented for a test case, highlighting the differences with a computational fluid dynamics and flight dynamics full–coupled model. A blocks software architecture is used to obtain a tool open and customizable.

A computational fluid dynamics based optimization loop is then used to analyse the longitudinal trim conditions of a test case, presenting the derivatives aerodynamic model limitations. The geometry optimization feasibility, considering the aircraft stability as objective, is assessed. A model based on aerodynamic derivatives is assumed for the representation of the aerodynamic loads, because traditionally used by flight dynamics tools. Advances in this direction are discussed.

More information

Identifiers

ORCID for A. Da Ronch: orcid.org/0000-0001-7428-6935

Catalogue record

Export record