This thesis focuses on the use of LES to simulate the flow around elliptical bluff body with blunt trailing edge fitted with open base cavity. The main objective of this study is to determine the effects of the cavity on the drag of the body. A secondary but important objective is to demonstrate that LES can provide accurate representation of the flow around this bluff body. Moreover, LES results can complement the available experimental results in order to provide a much better understanding of the flow. The simulations were carried out at a Reynolds number of 2.6x104 based on the height of the body using Spalart-Allmaras RANS model while the LES were performed using Smagorinsky dynamic model. A grid-independence test was conducted using three grids which contain 0.85M, 1.3M and 1.7M cells, respectively. This test shows that the results are grid-independent. The LES results predicted the mean flow field in the near wake with good accuracy as compared to the experimental mean flow field obtained. The base pressure results show that the base pressure coefficient for the base model was around -0.56, which agrees well with the experimental results .By attaching the cavity, the base pressure has increased. The increase in base pressure coefficient was around 44% using 1/3 h cavity and this agrees well with the experimental measurements. The RANS predicted drag coefficient of 0.56 for the base model and 0.471 for the cavity model. This represents a difference of 8% for the base model and 34% for the cavity model when compared with experiment drag coefficients (0.61 for the base model and 0.35 for the cavity model). For the LES, the drag coefficient of the base model was around 0.65 (6.5% difference) and using the cavity, the drag coefficient was reduced to around 0.37 (5.74% difference).;Details of the mean velocity components have been compared with experimental data at various locations in the wake region of the flow. Observation on the comparison between LES and RANS shows that LES predicted the mean flow field more accurately than RANS particularly downstream the recirculation regions. The length of the recirculation region was over predicted by RANS compared to LES. The prediction of this length by LES was in excellent agreement with experimental measurement.
展开▼
