Demand for faster CFD simulation exist and it is going to exist. Quick turn around time is paramount in all design environments and it is true for CFD based design as well. In general, in design cycle, simplified configurations are studied to reduce the CFD design cycle time. CFD simulations are prohibitively costly for complex configurations, unsteady simulations, combustion studies and flow with high temperature effects.
CFD simulation time may be improved either in the CFD solver or in the associated computer hardware. Various numerical methods like local time stepping, implicit methods, residual smoothing, dual time stepping, multi-grid methods are available for to improve both robustness and convergence. Parallel clusters with multiple processing machines are also available to carry out simulations faster. Even with the advancement of high performance computing, only RANS is the industrial standard presently.
Therefore, in the present research proposal, a Non-Zonal Method will be investigated to improve the CFD simulation time. In the Zonal Method, one has to decide the zones and solver a-priori, like the viscous region near the body is solved using Navier-Stokes (NS) or Reynolds Averaged Navier-Stokes (RANS) solvers and in the regions away the body, Euler (inviscid) solver is used. Whereas, in the Non-Zonal Method, the solver fidelities automatically switches to an appropriate solver in a domain, thus saving the time required for CFD simulations. The main crux of the non-zonal method is to identify a appropriate switching parameters and its cut-off parameters to switch solver fidelities in a CFD domain for various flow speeds.