Implementation of aerodynamic identification technology based on the fastest descent method

Based on operator theory and numerical iteration technology, this paper quickly completes the identification of aerodynamic equivalent kernel functions with input-and-output data from the computational fluid dynamics code (CFD code). The convergence criterion of the numerical calculation is whether the error between identified forces and the given experimental data is acceptable, namely a rather accuracy. Compared with the classical impulse response method or step response method, our method does not attempt to directly identify the exact kernel functions of the system, but to approach to those approximation, namely, the equivalent kernel functions. Several examples are given to show this new technique to obtain their kernel functions, including a classic circuit system described by ordinal differential equations (Ode) and flow field analysis in the case with a variable Mach number. Once the system feature extraction is complete, the response is immediately obtained by the well-known convolution operator.