Matthieu Jeanneau

AWIATOR's Study of a Wing Load Control: Design and Flight-Test Results

Abstract Loads at the wing root sections of an aircraft are mainly induced by the vertical accelerations encountered in manoeuvre or in turbulence, but a non-negligible part comes from the wing bending deflections. In an attempt to reduce such loads and hence reduce structural weight, a load alleviation function has been studied by Airbus flight-control research-engineers. This study was conducted under the auspices of AWIATOR, a European research program aimed at developing and testing challenging flying technologies. The load alleviation function developed in this study is divided into 3 parts, each with a dedicated objective. First, a passive control reduces the loads induced by pilot inputs or by turbulence, by deflecting ailerons and spoilers proportionally to the vertical acceleration of the aircraft. Secondly, an active control deals with wing oscillations induced by the bending structural modes. This part has been designed with modern H methods, to take into account and optimise various specifications: load reduction, robustness to payload, but also roll-on and roll-off criteria to avoid any interaction possibly modifying handling qualities. The third part concerns the activation logics of the first two parts. This last part is not discussed hereafter.

Active Wing Load Alleviation with an Adaptive Feed- forward Control Algorithm

The latest subsonic civil transport aircrafts with high aspect ratio at low structural weight/payload ratio have been equipped with active wing-load control systems based on robust feedback of modal accelerations. The aim of these systems is to increase handling qualities and passenger comfort as well as to reduce dynamic wing-loads mainly induced by atmospheric disturbances. Structural control must be robust regarding variations in the plant transfer functions caused by the varying flight and load conditions to which an aircraft structure is exposed to. Such robustness criteria limit performance of robust wing-load control. Since robust structural control law design requires very accurate aero-elastic plant models, it is generally optimized iteratively during the flight test phase, where the accuracy of plant models is successively improved and the flight and load envelope is continuously expanded. This paper shows that introduction of an adaptive structural feed-forward control system could dramatically increase attainable performance of active wing-load control due to the feed-forward character on the one hand, and due to adaptivity on the other. The authors found that the best performance and robust stability of the adaptive wing-load alleviation system can be reached with an adaptive feed-forward controller in combination with error-feedback based on a stochastic-gradient-descent algorithm. The proposed algorithm rapidly adapts to any changes in plant transfer functions and excitation, so that extra robust stability margins no longer have to be taken into account. Additionally, the plant transfer functions have to be known in advance only approximately, since the controller optimizes itself online to the actual plant transfer function. Iterative design with the help of flight test optimized plant models is no longer necessary to perform. Thus it is estimated that the design period and design costs for active wing-load control can be dramatically decreased. After introducing the basic principle and integration of the proposed adaptive controller, the mathematical background of the algorithm is briefly discussed. This is followed by a detailed stability analysis to ensure stable adaptation of the algorithm. Finally, results of numeric simulations are presented to underline the validity of found stability conditions and to highlight improved performance, compared to a robust structural feedback control system.

The AIRBUS On-Ground Transport Aircraft Benchmark

This chapter describes the behaviour of a transport aircraft and its systems during rolling. Notations and conventions are given first, followed by the main equations of motion. Loads affecting aircraft motion are then described and their modelling given. Finally a short aircraft behaviour analysis is provided. This chapter aims at offering the reader a clear understanding of the control application and its requirements.

AIRCRAFT MODELLING FOR NONLINEAR AND ROBUST CONTROL DESIGN AND ANALYSIS

Abstract In this paper an application of an exact nonlinear symbolic LFT modelling approach to an On-Ground Airbus aircraft is shown. The modelling approach used combines the natural modularity and clarity of presentation from LFT modelling with the ease of manipulation from symbolic algebra. It results in an exact nonlinear symbolic LFT that represents an ideal starting point to perform subsequent simplifications and assumptions to finally transform the model into an approximated symbolic LFT ready for design and analysis.

AWIATOR'S DESIGN OF MULTI-OBJECTIVES CONTROL LAWS

Abstract Following the works presented in St Petersburg during the 16th IFAC Symposium on Automatic Control in Aerospace, Airbus has been working on the design and evaluation of more integrated control laws, taking into account multi-objectives design criteria, gathering from the very beginning the specifications in term of handling qualities, comfort, loads, and robustness. A non-conservative H ∞ schema allows to directly associate each criterion to a dedicated exogenous transfer function and a tuning weighting. This approach offers an easy tuning, criterion by criterion, guaranteeing optimal results. This is also very convenient for industrial evaluation of reachable compromises between criteria, and Pareto-like plotting.

Adaptive Output Feedback Control Based on Neural Networks: Application to Flexible Aircraft Control

Abstract Abstract One of the major challenges in aeronautical flexible structures control is the uncertain or the non stationary feature of the systems. Transport aircrafts are of unceasingly growing size but are made from increasingly light materials so that their motion dynamics present some flexible low frequency modes coupled to rigid modes. For reasons that range from fuel transfer to random flying conditions, the parameters of these planes may be subject to significative variations during a flight. A single control law that would be robust to so large levels of uncertainties is likely to be limited in performance. For that reason, we follow in this work an adaptive control approach. Given an existing closed-loop system where a basic controller controls the rigid body modes, the problem of interest consists in designing an adaptive controller that could deal with the flexible modes of the system in such a way that the performance of the first controller is not deteriorated even in the presence of parameter variations. To this purpose, we follow a similar strategy as in Hovakimyan (2002) where a reference model adaptive control method has been proposed. The basic model of the rigid modes is regarded as a reference model and a neural network based learning algorithm is used to compensate online for the effects of unmodelled dynamics and parameter variations. We then successfully apply this control policy to the control of an Airbus aircraft. This is a very high dimensional dynamical model (about 200 states) whose direct control is obviously hard. However, by applying the aforementioned adaptive control technique to it, some promising simulation results can be achieved.

Dynamic flight control laws, real-time implementation issues

Abstract The works presented in this paper focus on implementation issues of complex dynamic flight control laws. Ambitious multi-criteria objectives and the need to cover the full flight domain of a transport aircraft increase implementation difficulties: interpolation of the control laws, numerical and discretization issues and fast retiming. Relevant methods for MIMO dynamic control laws are proposed, which proved their efficiency on both lateral and longitudinal aircraft control.

Neural Networks Contribution to Modeling for Flight Control

Today, civil aviation is facing new challenges in nonlinear flight control design. Recent nonlinear control techniques offer solutions to these challenges but also bring the need for onboard models of numerical aerodynamics coefficients. The requirements on these potentially onboard models are very strong, since they must be accurate, reliable and compact to cope with aeronautical designs golden rules. It appears that neural networks can meet the aeronautical requirements. However, the usual neural networks design tools are neither autonomous nor fast enough for standard industrial use. We developed integrated neural network identification software to create new automated tools needed for aeronautical industrial applications, such as architecture optimisation and maximum statistical error quantification.

On-Ground Transport Aircraft Nonlinear Control Design and Analysis Challenges

This chapter provides the main guidelines for the control design and analysis of a “rolling on the ground” control law. The rolling of aircraft is a very challenging task in term of piloting. The overall design of transport aircraft is clearly not optimized for rolling on ground but for flight. Its natural rolling qualities are very poor, both in term of stability and performance. Besides the coupling of aerodynamic loads, engine-thrusts, gravity and friction loads at the contact point of each tyre produce highly nonlinear and time-varying dynamics. These dynamics are strongly influenced by many parameters such as the velocity, the runway state, aircraft configuration (mass and inertia) and external disturbances like wind turbulence or gusts. The control objectives may also vary depending on the rolling phase: taxiing, runway acceleration prior to take-off, runway deceleration after landing, and runway deceleration after a rejected take-off.

CONTROL LAWS FOR AUTOMATIC TAKE-OFF USING ROBUST NONLINEAR INVERSION TECHNIQUES

Abstract This paper describes a fully automatic take-off function currently developed by Airbus for its airliners. The take-off phase is a very challenging problem in term of control due to nonlinearities occurring during both acceleration and rotation. Besides this is a phase with fast time-varying dynamics. Main actions to do during a take-off are recalled. The newly automatic take-off mode is then described. A focus on the pitch axis control is provided, using the nonlinear RMI control design technique. Results illustrate the very good behaviour of the overall function.