2D inverse convection dominated problem for estimation of inflow parameters from outflow measurements

Abstract

The estimation of inflow parameters from outflow measurements is considered in the optimization statement for two-dimensional supersonic laminar flow governed by the parabolized Navier-Stokes equations. The discrepancy of measured and calculated data on outflow boundary is minimized by variation of inflow parameters using the gradient-based methods. The discrepancy gradient is calculated using the adjoint problem. Both forward and adjoint problems are solved by the finite difference method. The numerical tests demonstrate the instability, especially in presence of data error. This instability may be handled using either the discrepancy principle or second order regularizing terms. Numerical tests demonstrate the feasibility of inflow parameter estimation.

Keywords:

• Adjoint problem
• Parabolized Navier-Stokes equations
• Parameter estimation
• Gradient optimization

^* [email protected]

^* [email protected]

Notes

^* [email protected]