The large eddy simulation (LES) of compressible transitional flows in a low-pressure turbine cascade is performed by using 6 th -order compact difference and 10 th -order filtering method. The numerical results without free-stream turbulence and those with about 5% of free-stream turbulence are compared. In these simulations, separated-flows in the turbine cascade accompanied by laminar-turbulent transition are realized, and the present results closely agree with past experimental measurements in terms of the static pressure distribution around the blade. In the case where no free-stream turbulence is taken into account, unsteady pressure field essentially differs from that with strong free-stream turbulence. In the no free-stream turbulence case, pressure waves that propagate from blade's wake region have noticeable effects on the separated-boundary layer near the trailing-edge, and on the neighboring blade. Also, based on Snapshot Proper Orthogonal Decomposition (POD) analysis, dominant behaviors of the transitional boundary layers are investigated.

Effects of circulation on the evolution of vortex tubes and the associated response of near-wall flows in the shear of laminar boundary-layer flows are investigated using a model proposed by Hon and Walker (Hon, T.L. & Walker, J.D.A, Computers & Fluids, 20(3), pp. 343-358, 1991). Direct numerical simulations with freestream Mach number of 0.5 are conducted. Firstly, the dynamics of single hairpin vortex is investigated. Numerous secondary hairpin vortices, much more than previously reported, which are regularly aligned in the streamwise direction are allowed to be newly generated according to the shear-layer instability of the legs of an initial hairpin vortex. Small-scale turbulence is then produced when the circulation is sufficiently large. Secondly, a straight vortex tube model is investigated. Sinuous deformation of a shear layer, which leads to the generation of discrete hairpin vortices, becomes obvious especially near the upper region of the vortex tube. In order to quantify the initial instability triggering the generation of the secondary hairpin vortices, quasi-linear stability analysis is conducted. While only one unstable mode appears when the circulation is small, two modes, that is, off-wall mode and near-wall mode, appear when the circulation is large. The cases of circulation where the two modes appear correspond to those of circulation where the production of small-scale turbulence is observed in the simulations of the single hairpin vortex.

Bends widely used in process piping systems can cause strong pressure fluctuations on pipe wall for a high-velocity flow, and hence, flow induced vibration (FIV) of piping occurs. Currently, the FIV assessment is made primarily based on the guideline published by Energy Institute. However, it is based on very conservative assumptions, and thus, results in excessive design of piping systems. The coupling analysis of CFD/FEA (Computational Fluid Dynamics/Finite Element Analysis) is expected to be a useful approach for more proper FIV assessment. The present study mainly aims at verifying CFD prediction accuracy of wall pressure fluctuations or FIV loadings around a pipe bend. In CFD benchmark study, large eddy simulations (LES) with dynamic Smagorinsky model (DSM) were performed for a 90° mitred bend used in the experiments in literature, under two different flow velocity conditions. The benchmark simulation results show that the power spectral density (PSD) of the LES-predicted wall pressure fluctuations at the sampling locations is near to the experimental results with moderate conservativeness desirable for engineering applications. Also, the LES-predicted peak frequencies are close to the experimental data. Therefore, it is suggested that the applied numerical approaches be applicable to predict the FIV loadings with moderately high accuracy for engineering applications.

When a severe accident occurs, decommissioning work becomes important task. In the decommissioning work after the severe accident, establishing the way to estimate the sedimentation place of molten debris is important. However, the technique to estimate exactly sedimentation place has not been enough. Therefore, the detailed and phenomenological numerical simulation code named JUPITER for predicting the molten core behavior is under development. The comparison between experimental and numerical results is necessary to clarify the validity of the numerical analysis code. This study provides the experimental data for a BWR to examine the numerical simulation code in order to contribute to progress of the decommissioning work.