Active flutter suppression for a three surface transport aircraft by recurrent neural networks

approach for the design of a flutter suppression system by means of Recurrent Neural Networks (RNNs). The controller is used to move flutter instabilities outside the flight envelope of an unconventional three surface, transport aircraft configuration. The design process requires a comprehensive aircraft model, where flight mechanics, structural dynamics, unsteady aerodynamics and control surface actuators are represented in state-space form, according to the “modern” aeroelastic approach. The control system implemented for flutter suppression is based on two RNNs: one is trained to identify system dynamics; the other works as a controller using an indirect inversion of the identified model. Keeping the training of both RNNs “on line” leads to an adaptive control system. Extensive numerical tests are used to tune the neural network design parameters and to show how the neural controller increases system damping, widening the flutter-free flight envelope by more than 15% of the uncontrolled flutter velocity. lutter can often be a critical problem for current flight vehicle design, due to the increase of structure lightness which leads to high structural flexibility. Several approaches may be followed to solve this problem. The first, which may be denominated “passive”, basically goes through a re-design of the aircraft to increase its stability boundaries. The second, which may be called “active”, tries to exploit the capabilities of control systems to improve aircraft stability properties, usually with a smaller weight increment. Furthermore, such controllers can be used to improve the performances and the cruise comfort. Here, an active control strategy based on Recurrent Neural Networks (RNNs) for flutter suppression is designed to improve the stability boundaries and performances of a transport aircraft with an unconventional configuration, denominated X-DIA. The X-DIA, sketched in figure 1, is a conceptual design of a short range 70-seats jet liner based on the idea of using three main lifting surfaces: a front canard, a 15 deg. forward swept wing and a T-tail. The rear location of the main wing along the fuselage, allowed by the forward swept wing coupled with the canard surfaces, is expected to yield a positive cooperation between structures and aerodynamics, allowing a significant weight saving and drag reduction. This configuration is currently the object of numerous investigations at the Dept. of Aerospace Engineering of Politecnico di Milano (DIAPM), both numerical and experimental. 1‐3 A 1/10, Froude scaled, wind tunnel model has been built and tests

[1]  H. E. Merritt,et al.  Hydraulic Control Systems , 1991 .

[2]  D. Sorensen,et al.  Approximation of large-scale dynamical systems: an overview , 2004 .

[3]  D. Bruce Owens,et al.  Exploratory wind tunnel investigation of the stability and control characteristics of a three-surface, forward-swept wing advanced turboprop model , 1990 .

[4]  Kumpati S. Narendra,et al.  Identification and control , 1998 .

[5]  K. L. Roger,et al.  Airplane Math Modeling Methods for Active Control Design , 1977 .

[6]  Paolo Mantegazza,et al.  Active Flutter Suppression Using Recurrent Neural Networks , 2000 .

[7]  J. W. Agnew,et al.  Benefits of Aerodynamic Interaction to the Three-Surface Configuration , 1980 .

[8]  Rolf Isermann,et al.  Adaptive control systems , 1991 .

[9]  W. Heinze,et al.  Three Surface Aircraft - A Concept For Future Transport Aircraft , 2000 .

[10]  W. P. Rodden,et al.  Equations of motion of a quasisteady flight vehicle utilizing restrained static aeroelastic characteristics , 1985 .

[11]  Pam Haley,et al.  Generalized predictive control for active flutter suppression , 1997 .

[12]  Paolo Mantegazza,et al.  Single Finite States Modeling of Aerodynamic Forces Related to Structural Motions and Gusts , 1999 .

[13]  Liang Jin,et al.  Adaptive control of discrete-time nonlinear systems using recurrent neural networks , 1994 .

[14]  E. Kendall.  The aerodynamics of three-surface airplanes , 1984 .

[15]  Graham C. Goodwin,et al.  Control System Design , 2000 .

[16]  W. Rodden,et al.  A doublet-lattice method for calculating lift distributions on oscillating surfaces in subsonic flows. , 1969 .

[17]  Robert C. Scott,et al.  Active control of wind-tunnel model aeroelastic response using neural networks , 2000, Smart Structures.

[18]  Kumpati S. Narendra,et al.  Control of nonlinear dynamical systems using neural networks: controllability and stabilization , 1993, IEEE Trans. Neural Networks.

[19]  Ronald J. Williams,et al.  A Learning Algorithm for Continually Running Fully Recurrent Neural Networks , 1989, Neural Computation.

[20]  Tommy W. S. Chow,et al.  A recurrent neural-network-based real-time learning control strategy applying to nonlinear systems with unknown dynamics , 1998, IEEE Trans. Ind. Electron..

[21]  Sergio Ricci,et al.  Active Control of Three Surface Wind Tunnel Aeroelastic Demonstrator: Modeling and Correlation , 2005 .

[22]  Fu-Chuang Chen,et al.  Adaptive control of nonlinear systems using neural networks , 1992 .

[23]  Jiri Cecrdle,et al.  Experimental investigations of a vibration suppression system for a three surface aeroelastic model , 2005 .

[24]  Sergio Ricci,et al.  Control of an all-movable foreplane for a three surfaces aircraft wind tunnel model , 2006 .

[25]  Franco Bernelli-Zazzera,et al.  Adaptive Control of Space Structures via Recurrent Neural Networks , 1999 .

[26]  Gregory J. Hancock.  An introduction to the flight dynamics of rigid aeroplanes , 1995 .

[27]  Bernard Friedland,et al.  Control System Design: An Introduction to State-Space Methods , 1987 .

[28]  Sergio Ricci,et al.  Active Aeroelastic Control Over a Multisurface Wing: Modeling and Wind-Tunnel Testing , 2007 .

[29]  K S Narendra,et al.  Control of nonlinear dynamical systems using neural networks. II. Observability, identification, and control , 1996, IEEE Trans. Neural Networks.

[30]  Bernard Etkin,et al.  Dynamics of Atmospheric Flight , 1972 .

[31]  L. Morino,et al.  Matrix fraction approach for finite-state aerodynamic modeling , 1995 .