Improved Boosting Model for Unsteady Nonlinear Aerodynamics based on Computational Intelligence

Improved Boosting Model for Unsteady Nonlinear Aerodynamics based on Computational Intelligence

Boxu Zhao, Guiming Luo, Jihong Zhu
DOI: 10.4018/IJCINI.2017010104
OnDemand:
(Individual Articles)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

The large-amplitude-oscillation experiment was carried out with two levels of freedom to provide data. Based on the wind tunnel data, polynomial regression, least-square support vector machines and radial basis function neural networks are studied and compared in this paper. An improved model was also developed in this work for unsteady nonlinear aerodynamics on the basis of standard boosting approach. The results on the wind tunnel data show that the predictions of the method are almost consistent with the actual data, thus demonstrating that these methods can model highly nonlinear aerodynamics. The results also indicate that improved boosting model has better accuracy than the other methods.
Article Preview
Top

1. Introduction

In current days, the growing requirements for the performance of contemporary fighter aircraft are the capability to implement stall maneuvers, including the capability to achieve steerable tactical maneuvers after the stalling angle of attack is exceeded by the aircraft. An aircraft is required to extend the flight envelope into an area with a large attack angle in order to accomplish a stall maneuver. (Liu, 2005) The status of flight greatly changes when an aircraft maneuver is in a region with a large angle of amplitude and attack. (Sun, 2011) Vortex breakdown, vortex flow, and flow separation might quickly show up in the flow nearby the aircraft. Currently, unsteady aerodynamic characteristics cannot be described by dynamic derivative models and an unsteady hysteresis effect and a strong nonlinear property are shown by the aerodynamic forces acting on the aircraft. Thus, experiment or research on a large angle of amplitude and attack is required. (Gong, 2007) In fact, nonlinear unsteady aerodynamics which has been accurately predicted plays an essential role in the analysis of the flight status of an aircraft to design the systems of flight control. The first requirement in the mathematical modeling of aircraft aerodynamics is the model’s structure, including determination of the expression, the algorithm, the preferred requirement and then the estimated parameters of the pneumatic response. Today, many aerodynamic modeling approaches have been put forward.

In 1911, Bryan proposed the aerodynamic coefficient linear model which developed the nonlinear algebraic models for a large angle of attack. This kind of model mainly includes the spline-function model and the polynomial model (Shi, 1999). Tobak in the 1970s utilized the step-response approach to build the nonlinear unsteady aerodynamic model in the integral form, generally expressing the unstable nonlinear aerodynamic forces. The state-space model was put forward by Goman in the beginning of the 1980s, emphasizing on the hysteresis effect incurred by vortex crushing and separation flow. (Chin, 1992) During that period of time, many approaches for the estimation of parameters were put forward, for instance, the maximum likelihood approach, the generalized Calman filtering algorithm, and the least-squares approach. Nowadays, a new means to design aircraft is offered by artificial intelligence and Computational intelligence. Modeling approaches, like artificial neural networks or fuzzy logic, with strong learning ability and high adaptability could approximate any nonlinear function in IJCINI.2017010104.m01. Therefore, artificial intelligence and computational intelligence have become very popular in the research of nonlinear unsteady aerodynamic modeling recently.

Complete Article List

Search this Journal:
Reset
Volume 18: 1 Issue (2024)
Volume 17: 1 Issue (2023)
Volume 16: 1 Issue (2022)
Volume 15: 4 Issues (2021)
Volume 14: 4 Issues (2020)
Volume 13: 4 Issues (2019)
Volume 12: 4 Issues (2018)
Volume 11: 4 Issues (2017)
Volume 10: 4 Issues (2016)
Volume 9: 4 Issues (2015)
Volume 8: 4 Issues (2014)
Volume 7: 4 Issues (2013)
Volume 6: 4 Issues (2012)
Volume 5: 4 Issues (2011)
Volume 4: 4 Issues (2010)
Volume 3: 4 Issues (2009)
Volume 2: 4 Issues (2008)
Volume 1: 4 Issues (2007)
View Complete Journal Contents Listing