Numerical simulation of transonic flows by a flexible optimiser evolution agent

Abstract

Our methodology is to use evolutionary algorithms (EAs) as a simulation and meshless method, because of the equivalence between verifying a numerical scheme for a nodal point and minimizing a corresponding objective function at this point. One advantage and powerful characteristic of our methodology is that mathematical knowledge and theories related to the problem can be efficiently and easily incorporated. The methodology proposed opens the possibility of solving the problem from a more local point of view as a domain decomposition problem. In its practical application, our aim is to obtain a good approximate solution of the exact solution, and from the good and useful information thereby made available, we can use a traditional simulation method to refine results, but starting the application of the method in a more appropriated way, for example, starting with an appropriate mesh. In this paper we continue our open research line. Here we get numerical results when a divergent nozzle has supersonic upstream flow at the entrance and subsonic flow at the exit. The steady state solution contains a transonic shock. The results agree with the exact position and conditions of the shock line.