“…Reduction by projection on a reduced space is often used in the literature because of its ease of implementation, but also because it keeps the general expression of the equation of motion so that the same integration algorithms as those used for the full order model can be used. Different reduction bases have been recently investigated in a nonlinear context by different authors: linear bases [9,10], bases obtained from Proper Orthogonal Decomposition (POD) [11,8], linear bases augmented with modal derivatives (MD) [12,13,14] and bases composed of nonlinear complex modes [15,16].…”