Liquid droplet impingement erosion (LDI) occurs in fluid machinery on which many droplets impinge at high speed. In the process of LDI erosion, the material surface becomes completely roughened by erosion pits and these pits get locally deeper and deeper even though the droplets uniformly impinge on the material surface. In this study, to clarify the mechanism of erosion progressing in the form of locally deep depression, the numerical analysis of droplet impingement on the deep pit filled with water is conducted. From this result, the Rayleigh waves generated in the material by the droplet impingement propagate on the surface of the pit and interfere at the bottom. Therefore the stress generated at the bottom is larger than that of the droplet impingement. This may explain why the deep erosion pits are formed in liquid droplet impingement erosion even though the many droplets uniformly impinge on the entire material surface.

This study deals with an attempt to optimize blade profiles of 2-stage axial turbine section for a rocket engine turbopump, where the turbine section was originally designed by JAXA (Japan Aerospace Exploration Agency) in the project called DDT (Dynamic Design Team). An optimization based genetic algoriothm in cooperation with the surrogate-model constructed from RANS (Reynolds-Averaged Navier-Stokes' Eqs) simulations is applied to the blade profile exploration under two-dimensional flow condition to achieve the stage efficiency maximization, along with blade shape morphing. Out of the all calculated combinations of blades profiles for 1R (First Stage Rotor), 2S (Second Stage Stator) and 2R (Second Stage Rotor), top-30 highly efficient combinations are then picked up and three-dimensionalized so that each three-dimensional (3D) stage efficiency is calculated in order to check whether the highly efficient combinations via 2D optimization still perform good in 3D situation.