Nozzle flows are important issues in aerodynamics, for example, flows from jet engines, rockets, etc. Before our investigation, a brief account of the novel finite volume (FV) methods we used is first given. These includes an upwind (ETAU) scheme and a centered (CESE) scheme. They are nominally 2nd order accurate in space and time, but possess many nice features that are needed in large scale complex flow computations, such as high fidelity, low dissipation, implicit LES, highly stable, nearly all speed, using unstructured grid, with an extremely simple NRBC (nonreflecting boundary condition), ...,etc.
  Numerous examples are then presented. They were topics in aerospace research. Without exception, all the numerical results are compared to their corresponding experimental data. These nozzle flows involve complicated flow phenomena, such as vortex roll-up, shock-vortex interaction, generation and propagation of acoustic waves, feedback loop, etc. Some of the examples were never tried before or failed by using other methods. We thus demonstrate that even a low order (2nd order) FV scheme can be used as a viable tool in the computation of complicated compressible flows.