In English

Potential of Rolls‐Royce HYDRA Code for Ship Hull Performance Prediction

Gabriel Östling ; Erik Wiberger
Göteborg : Chalmers tekniska högskola, 2011. 98 s. Report. X - Department of Shipping and Marine Technology, Chalmers University of Technology, Göteborg, Sweden; 248, 2011.
[Examensarbete på avancerad nivå]

In this study the potential of Rolls‐Royce in‐house CFD code HYDRA, designed for high speed flows in gaseous fluids, is investigated for ship performance prediction in a marine environment. Using two different test cases in form of well known Wigley hull shapes the code is validated against commercially available software FLUENT and previously known results from experimental testing. Since HYDRA requires an ideal gas as the working fluid the surrounding conditions have been scaled from water to a fictional ideal gas with respect to the Reynolds number. Free surface solution is not possible at this point in HYDRA, therefore the hulls have been considered fully submerged. The parameters that have been studied are Mach number, mesh, mesh boundary layer resolution, convergence and CPU time. For comparison of the results the total resistance, as well as its decomposition into pressure and viscous resistance, have been observed. Simulations have been done at Mach numbers between 1 · 10􀬿􀬺 and 0.7 and 􀝕􀬾 between 1 and 120 using both k‐ω SST and Spalart‐Allmaras as turbulence models. Both structured and unstructured meshes have been applied. The results show that HYDRA, using an unstructured mesh with Mach number 0.01 and 􀝕􀬾 at 1, calculates the total resistance 2.3% higher than FLUENT and 3.85% lower than experimental results. However, as the Mach number in HYDRA increases the results become inconsistent, e.g. between Mach number 0.01 and 0.25 the total resistance increases around 3%. Even if that would bring HYDRA closer to the experimental results it is a highly unwanted effect since consistency is considered significantly more important than accuracy in CFD. When using unstructured meshes with higher 􀝕􀬾 HYDRA displays difficulties predicting accurate viscous and pressure resistance contributions. With a structured Mesh HYDRA is comparable to FLUENT and shows none of the above mentioned problems. In CPU time FLUENT was proven to be superior for these type of cases.

Nyckelord: Computational Fluid Dynamics, Wigley hull, FLUENT, HYDRA, GAMBIT, RANS, Spalart‐Allmaras,

Publikationen registrerades 2011-12-30. Den ändrades senast 2013-04-04

CPL ID: 151285

Detta är en tjänst från Chalmers bibliotek