Dynamical Medium (Large)- Scale Model Reduction and Interpolation with Application to Aircraft Systems
C. Poussot-Vassal, F. Demourant
Although the need for even more accurate system, phenomena and process
modeling is required in order to reduce development time and costs, the
number of variables linear and non-linear optimization tools can handle is still
a practical and theoretical limiting factor. This is especially true in aircraft
dynamical performance analysis, monitoring and control design, where dynamical models are accurately designed at varying local flight configurations, in order to handle flexible modes, aerodynamic delays, etc., leading to high-dimensional problems . Although Onera has a well established tradition of proposing complete and efficient tools for optimizing controllers and analyzing dynamic system performances through the use of Linear Fractional Representation (LFR) mathematical objects [2, 15, 22], recent growth in the dimensions of models has led to strong time and computational limitations when using these tools. The aim of this paper is to give an overview of the solutions developed within Onera to approximate a set of large-scale dynamical models with a parameterized LFR lower order model, which can be used in place of the original ones to effectively synthesize control laws and achieve performance analysis.