Identification of subsonic jet by dynamical recursive neural network and validation by Lyapunov exponents

The goal of this article is to identify a subsonic jet model generated in an industrial wind tunnel (blower) from actual measurements of the speed. For that, a new phase space is constructed by delay method (MOD). This construction enables one to extract invariants from the system. The identification based on a dynamical recursive neural network uses the time series of measurement. The validation of the model is achieved if the invariants resulting from the real system and those resulting from the model are nearly identical