This paper presents a method to obtain a Linear Fractional Transformation (LFT) model from a set of numerical flexible aircraft models. The core of the method lies in a polynomial interpolation, but before this step, some cautious treatments must be applied to the numerical models for them to be interpolable. First the models have to be reduced to be exploitable, while ensuring they all keep the same consistent modal content and avoiding numerical problems. Second the interpolation of the reduced models state representations can only be processed if the state vector has the same physical meaning for the whole model set (state vector consistency). For this problem, a specific state coordinate transformation is emphasized. After models are interpolated, the LFT is computed with the LFR toolbox's generalized Morton method. Then the LFT accuracy can be improved through a biconvex optimization.
展开▼
