Continuous and discontinous compressible flows in a converging-diverging channel solved by physics-informed neural networks without data

TL;DR

This paper demonstrates that physics-informed neural networks can accurately solve complex compressible flow problems in a converging-diverging nozzle without data, capturing both continuous and discontinuous solutions including shock phenomena.

Contribution

It introduces a novel approach to adjust PINNs for solving hyperbolic PDEs in compressible flows without relying on external data, successfully capturing flow branches and shocks.

Findings

PINNs can solve steady and unsteady compressible flows in nozzles.

The method captures discontinuous shock solutions accurately.

Different flow regimes are obtained by varying pressure ratios.

Abstract

Physics-informed neural networks (PINNs) are employed to solve the classical compressible flow problem in a converging-diverging nozzle. This problem represents a typical example described by the Euler equations, thorough understanding of which serves as a guide for solving more general compressible flows. Given a geometry of the channel, analytical solutions for the steady states indeed exist and they depend on the ratio between the back pressure of the outlet and stagnation pressure of the inlet. Moreover, in the diverging region, the solution may branch into subsonic flow, supersonic flow, and a mixture of both with a discontinuous transition where a normal shock takes place. Classical numerical schemes with shock-fitting/capturing methods have been designed to solve this type of problem effectively, whereas the original PINNs fail in front of the hyperbolic non-linear partial differential equations. We make a first attempt to exploit the power of PINNs to directly solve this problem by adjusting the weights of different components of the loss function, to acquire physical solutions and meanwhile avoid trivial solutions. With a universal setting yet no exogenous data, we are able to solve this problem accurately, that is, for different given pressure ratios PINNs provide different branches of solutions at both steady and unsteady states, some of which are discontinuous in nature.